diff --git a/src/common/snippets/include/snippets/lowered/port_descriptor.hpp b/src/common/snippets/include/snippets/lowered/port_descriptor.hpp
index 3fc429bec4df1e..926136abf0e8f8 100644
--- a/src/common/snippets/include/snippets/lowered/port_descriptor.hpp
+++ b/src/common/snippets/include/snippets/lowered/port_descriptor.hpp
@@ -54,6 +54,9 @@ class PortDescriptor {
     void set_reg_type(RegType type) { m_reg.type = type; }
     void set_reg_idx(size_t idx) { m_reg.idx = idx; }
 
+    // Indexing starts from the end (rbegin() + idx)
+    void set_subtensor_value(size_t idx, VectorDims::value_type value);
+
     std::string serialize() const;
     bool empty() const { return m_layout.empty() && m_subtensor_shape.empty();}
     PortDescriptorPtr clone() const;
diff --git a/src/common/snippets/src/lowered/pass/propagate_subtensors.cpp b/src/common/snippets/src/lowered/pass/propagate_subtensors.cpp
index 168bcf522585a1..ef64c4205b901a 100644
--- a/src/common/snippets/src/lowered/pass/propagate_subtensors.cpp
+++ b/src/common/snippets/src/lowered/pass/propagate_subtensors.cpp
@@ -18,11 +18,11 @@ namespace {
 
 // The algorithm uses the following special values in subtensors/shapes:
 // 1. Dynamic value in subtensor/shape : SIZE_MAX
-// 2. Full fimension in subtensor      : SIZE_MAX - 1
+// 2. Full dimension in subtensor      : SIZE_MAX - 1
 // 3. Default value of `new_dim_value` : SIZE_MAX - 2
 // 4. `Forced` special dynamic value   : SIZE_MAX - 3
 //
-// We have to introduce `SPECIAL_DYNAMIC_VALUE` to distinguish `new_dim_value = DYNAMIC`
+// We have to introduce `FORCED_DYNAMIC_VALUE` to distinguish `new_dim_value = DYNAMIC`
 // from the real dynamic values in subtensors and shapes and force this value in subtensors.
 // For example, there is Brgemm with the following info in the tail Loop:
 // Input 0: shape [?, ?], existing subtensor [32, FULL_DIM]
@@ -36,7 +36,7 @@ namespace {
 // 3. Update subtensor on output using shape:
 //    new_subtensor[i] = std::min(planar_shape[i], subtensor[i]); // i = 0: std::min(SIZE_MAX(?), 32)
 //    new subtensor [32, FULL_DIM] - has not been changed! But should be [?, FULL_DIM]
-// Conculsion: we have to distinguish forced dynamic value with existing dynamic values in shape and subtensor
+// Conclusion: we have to distinguish forced dynamic value with existing dynamic values in shape and subtensor
 
 constexpr size_t NEW_DEFAULT_VALUE    = SIZE_MAX - 2;
 constexpr size_t FORCED_DYNAMIC_VALUE = SIZE_MAX - 3;
@@ -61,9 +61,8 @@ void propagate_updated_subtensor_through_loop(const LinearIR& linear_ir,
                 const auto& expr = port.expr_port->get_expr();
                 const auto& desc = port.expr_port->get_descriptor_ptr();
                 auto subtensor = desc->get_subtensor();
-                if (port.dim_idx < subtensor.size()) {
-                    *(subtensor.rbegin() + port.dim_idx) = new_dim_value;
-                    desc->set_subtensor(subtensor);
+                if (port.dim_idx < desc->get_subtensor().size()) {
+                    desc->set_subtensor_value(port.dim_idx, new_dim_value);
                 }
 
                 const auto parent_desc = expr->get_input_port_connector(port.expr_port->get_index())->get_source().get_descriptor_ptr();
diff --git a/src/common/snippets/src/lowered/port_descriptor.cpp b/src/common/snippets/src/lowered/port_descriptor.cpp
index 522f03367f51c1..979f3120a68476 100644
--- a/src/common/snippets/src/lowered/port_descriptor.cpp
+++ b/src/common/snippets/src/lowered/port_descriptor.cpp
@@ -54,6 +54,11 @@ void PortDescriptor::set_shape(const VectorDims& tensor) {
     *m_tensor_shape = tensor;
 }
 
+void PortDescriptor::set_subtensor_value(size_t idx, VectorDims::value_type value) {
+    OPENVINO_ASSERT(idx < m_subtensor_shape.size(), "Failed to set subtensor value: idx should be less than size");
+    *(m_subtensor_shape.rbegin() + idx) = value;
+}
+
 PortDescriptorPtr PortDescriptor::clone() const {
     auto desc = std::make_shared<PortDescriptor>(*m_tensor_shape, m_subtensor_shape, m_layout);
     desc->set_reg(m_reg);
diff --git a/src/common/snippets/src/runtime_configurator.cpp b/src/common/snippets/src/runtime_configurator.cpp
index c0725c4f5b38aa..971e7a03858fcf 100644
--- a/src/common/snippets/src/runtime_configurator.cpp
+++ b/src/common/snippets/src/runtime_configurator.cpp
@@ -186,7 +186,7 @@ void RuntimeConfigurator::update_loop_info(const std::shared_ptr<lowered::Linear
         current_work_amount -= expanded_loop_info->get_work_amount();
 
         if (expanded_loop_info->is_evaluate_once()) {
-            expanded_loop_info->update_ptr_increments(std::vector<int64_t>(ptr_increments.size(), 0));
+            // Update only `finalization offsets`. `Ptr increments` are always zeroed in this case
             auto updated_finalization_offsets = current_work_amount > 0 ? std::vector<int64_t>(finalization_offsets.size(), 0) : finalization_offsets;
             // work_amount is equal to increment in cases with `evaluate_once`
             for (size_t i = 0; i < updated_finalization_offsets.size(); ++i)
diff --git a/src/common/snippets/tests/include/lir_test_utils.hpp b/src/common/snippets/tests/include/lir_test_utils.hpp
index ebe2e63cc2e66f..b653c86af8ab0b 100644
--- a/src/common/snippets/tests/include/lir_test_utils.hpp
+++ b/src/common/snippets/tests/include/lir_test_utils.hpp
@@ -44,40 +44,6 @@ void init_expr_descriptors(const ov::snippets::lowered::ExpressionPtr& expr,
                            const std::vector<ov::snippets::VectorDims>& subtensors = {},
                            const std::vector<ov::snippets::VectorDims>& layouts = {});
 
-/**
- * @brief Creates unified loop info based on provided entry and exit points, and adds it to the linear_ir's loops map
- * @attention This helper wraps LoopManager::mark_loop method, but only for LoopInfo creation (whereas original
- * mark_loop method also marks expressions with the corresponding loop info).
- * @param linear_ir linear_ir in which loop info should be added
- * @param entries entry points of loop
- * @param exits exit points of loop
- * @return ID of created loop
- */
-size_t create_and_add_unified_loop_info(const std::shared_ptr<ov::snippets::lowered::LinearIR>& linear_ir,
-                                        size_t work_amount,
-                                        size_t increment,
-                                        const std::vector<ov::snippets::lowered::LoopPort>& entries,
-                                        const std::vector<ov::snippets::lowered::LoopPort>& exits,
-                                        bool add_default_handlers = true);
-/**
- * @brief Creates unified loop info based on provided entry and exit points, and adds it to the linear_ir's loops map.
- * Meanwhile set loop id to expr range [loop_begin_pos, loop_end_pos).
- * @attention This helper wraps LoopManager::mark_loop method, which also marks expressions with the corresponding loop info
- * @param linear_ir linear_ir in which loop info should be added
- * @param loop_begin_pos begin expr postion in this loop
- * @param loop_end_pos end expr postion in this loop
- * @param entries entry points of loop
- * @param exits exit points of loop
- * @return ID of created loop
- */
-size_t create_and_add_unified_loop_info(const std::shared_ptr<ov::snippets::lowered::LinearIR>& linear_ir,
-                                        ov::snippets::lowered::LinearIR::constExprIt loop_begin_pos,
-                                        ov::snippets::lowered::LinearIR::constExprIt loop_end_pos,
-                                        size_t work_amount,
-                                        size_t increment,
-                                        const std::vector<ov::snippets::lowered::LoopPort>& entries,
-                                        const std::vector<ov::snippets::lowered::LoopPort>& exits,
-                                        bool add_default_handlers = true);
 }  // namespace snippets
 }  // namespace test
 }  // namespace ov
diff --git a/src/common/snippets/tests/src/lir_test_utils.cpp b/src/common/snippets/tests/src/lir_test_utils.cpp
index 9b6a1c8f84af69..084a5a34a3bd55 100644
--- a/src/common/snippets/tests/src/lir_test_utils.cpp
+++ b/src/common/snippets/tests/src/lir_test_utils.cpp
@@ -85,28 +85,6 @@ void init_expr_descriptors(const ov::snippets::lowered::ExpressionPtr& expr,
     }
 }
 
-size_t create_and_add_unified_loop_info(const LinearIRPtr& linear_ir,
-                                        size_t work_amount,
-                                        size_t increment,
-                                        const std::vector<LoopPort>& entries,
-                                        const std::vector<LoopPort>& exits,
-                                        bool set_default_handlers) {
-    // Equal begin and end iterators are set to avoid expressions marking with new loop id
-    return create_and_add_unified_loop_info(linear_ir, linear_ir->begin(), linear_ir->begin(), work_amount, increment, entries, exits, set_default_handlers);
-}
-
-size_t create_and_add_unified_loop_info(const LinearIRPtr& linear_ir,
-                                        ov::snippets::lowered::LinearIR::constExprIt loop_begin_pos,
-                                        ov::snippets::lowered::LinearIR::constExprIt loop_end_pos,
-                                        size_t work_amount,
-                                        size_t increment,
-                                        const std::vector<LoopPort>& entries,
-                                        const std::vector<LoopPort>& exits,
-                                        bool set_default_handlers) {
-    const auto& loop_manager = linear_ir->get_loop_manager();
-    return loop_manager->mark_loop(loop_begin_pos, loop_end_pos, work_amount, increment, entries, exits, set_default_handlers);
-}
-
 }  // namespace snippets
 }  // namespace test
 }  // namespace ov
diff --git a/src/common/snippets/tests/src/lowered/pass/extracted_loop_invariants.cpp b/src/common/snippets/tests/src/lowered/pass/extracted_loop_invariants.cpp
index c3f4f5ea7f6877..ee762f4bfca746 100644
--- a/src/common/snippets/tests/src/lowered/pass/extracted_loop_invariants.cpp
+++ b/src/common/snippets/tests/src/lowered/pass/extracted_loop_invariants.cpp
@@ -66,11 +66,11 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsWithParams) {
         auto result = linear_ir->push_node<ov::opset10::Result>(sub.second);
         auto begin = multiply.first;
         auto end = result.first;
-        create_and_add_unified_loop_info(linear_ir, begin, end, 512, vector_size,
-                                        {LoopPort((*multiply.first)->get_input_port(0)),
-                                         LoopPort((*multiply.first)->get_input_port(1)),
-                                         LoopPort((*sub.first)->get_input_port(0))},
-                                        {LoopPort((*sub.first)->get_output_port(0))});
+        linear_ir->get_loop_manager()->mark_loop(begin, end, 512, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0)),
+                                                                       LoopPort((*multiply.first)->get_input_port(1)),
+                                                                       LoopPort((*sub.first)->get_input_port(0))},
+                                                 std::vector<LoopPort>{LoopPort((*sub.first)->get_output_port(0))});
         linear_ir->set_loop_depth(1);
     }
     {
@@ -85,10 +85,10 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsWithParams) {
         auto result = linear_ir_ref->push_node<ov::opset10::Result>(sub.second);
         auto begin = sub.first;
         auto end = result.first;
-        create_and_add_unified_loop_info(linear_ir_ref, begin, end, 512, vector_size,
-                                        {LoopPort((*sub.first)->get_input_port(0)),
-                                         LoopPort((*sub.first)->get_input_port(1))},
-                                        {LoopPort((*sub.first)->get_output_port(0))});
+        linear_ir_ref->get_loop_manager()->mark_loop(begin, end, 512, vector_size,
+                                                     std::vector<LoopPort>{LoopPort((*sub.first)->get_input_port(0)),
+                                                                           LoopPort((*sub.first)->get_input_port(1))},
+                                                     std::vector<LoopPort>{LoopPort((*sub.first)->get_output_port(0))});
     }
 }
 
@@ -124,10 +124,10 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsWithScalar) {
         auto result = linear_ir->push_node<ov::opset10::Result>(sub.second);
         auto begin = scalar.first;
         auto end = result.first;
-        create_and_add_unified_loop_info(linear_ir, begin, end, 512, vector_size,
-                                        {LoopPort((*multiply.first)->get_input_port(0)),
-                                         LoopPort((*sub.first)->get_input_port(0))},
-                                        {LoopPort((*sub.first)->get_output_port(0))});
+        linear_ir->get_loop_manager()->mark_loop(begin, end, 512, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0)),
+                                                                       LoopPort((*sub.first)->get_input_port(0))},
+                                                 std::vector<LoopPort>{LoopPort((*sub.first)->get_output_port(0))});
         linear_ir->set_loop_depth(1);
     }
     {
@@ -142,10 +142,10 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsWithScalar) {
         auto result = linear_ir_ref->push_node<ov::opset10::Result>(sub.second);
         auto begin = sub.first;
         auto end = result.first;
-        create_and_add_unified_loop_info(linear_ir_ref, begin, end, 512, vector_size,
-                                        {LoopPort((*sub.first)->get_input_port(0)),
-                                         LoopPort((*sub.first)->get_input_port(1))},
-                                        {LoopPort((*sub.first)->get_output_port(0))});
+        linear_ir_ref->get_loop_manager()->mark_loop(begin, end, 512, vector_size,
+                                                     std::vector<LoopPort>{LoopPort((*sub.first)->get_input_port(0)),
+                                                                           LoopPort((*sub.first)->get_input_port(1))},
+                                                     std::vector<LoopPort>{LoopPort((*sub.first)->get_output_port(0))});
     }
 }
 
@@ -187,20 +187,20 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsOutputLoopUpdateNotNeed
         auto result1 = linear_ir->push_node<ov::opset10::Result>(sub.second);
         auto begin = multiply.first;
         auto end = result1.first;
-        create_and_add_unified_loop_info(linear_ir, begin, end, 16, vector_size,
-                                        {LoopPort((*multiply.first)->get_input_port(0), true, 0),
-                                         LoopPort((*multiply.first)->get_input_port(1), true, 0),
-                                         LoopPort((*add.first)->get_input_port(0), true, 0),
-                                         LoopPort((*sub.first)->get_input_port(0), true, 0)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 0),
-                                         LoopPort((*sub.first)->get_output_port(0), true, 0)});
-        create_and_add_unified_loop_info(linear_ir, begin, end, 3, 1,
-                                        {LoopPort((*multiply.first)->get_input_port(0), true, 1),
-                                         LoopPort((*multiply.first)->get_input_port(1), true, 1),
-                                         LoopPort((*add.first)->get_input_port(0), true, 1),
-                                         LoopPort((*sub.first)->get_input_port(0), true, 1)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 1),
-                                         LoopPort((*sub.first)->get_output_port(0), true, 1)});
+        linear_ir->get_loop_manager()->mark_loop(begin, end, 16, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*multiply.first)->get_input_port(1), true, 0),
+                                                                       LoopPort((*add.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*sub.first)->get_input_port(0), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 0),
+                                                                       LoopPort((*sub.first)->get_output_port(0), true, 0)});
+        linear_ir->get_loop_manager()->mark_loop(begin, end, 3, 1,
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0), true, 1),
+                                                                       LoopPort((*multiply.first)->get_input_port(1), true, 1),
+                                                                       LoopPort((*add.first)->get_input_port(0), true, 1),
+                                                                       LoopPort((*sub.first)->get_input_port(0), true, 1)},
+                                                 std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 1),
+                                                                       LoopPort((*sub.first)->get_output_port(0), true, 1)});
         linear_ir->set_loop_depth(2);
     }
     {
@@ -218,21 +218,21 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsOutputLoopUpdateNotNeed
         auto result1 = linear_ir_ref->push_node<ov::opset10::Result>(sub.second);
         auto begin_inner = add.first;
         auto end_inner = result1.first;
-        create_and_add_unified_loop_info(linear_ir_ref, begin_inner, end_inner, 16, vector_size,
-                                        {LoopPort((*add.first)->get_input_port(0), true, 0),
-                                         LoopPort((*add.first)->get_input_port(1), true, 0),
-                                         LoopPort((*sub.first)->get_input_port(0), true, 0)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 0),
-                                         LoopPort((*sub.first)->get_output_port(0), true, 0)});
+        linear_ir_ref->get_loop_manager()->mark_loop(begin_inner, end_inner, 16, vector_size,
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_input_port(0), true, 0),
+                                                                           LoopPort((*add.first)->get_input_port(1), true, 0),
+                                                                           LoopPort((*sub.first)->get_input_port(0), true, 0)},
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 0),
+                                                                           LoopPort((*sub.first)->get_output_port(0), true, 0)});
         auto begin_outer = multiply.first;
         auto end_outer = result1.first;
-        create_and_add_unified_loop_info(linear_ir_ref, begin_outer, end_outer, 3, 1,
-                                        {LoopPort((*multiply.first)->get_input_port(0), true, 1),
-                                         LoopPort((*multiply.first)->get_input_port(1), true, 1),
-                                         LoopPort((*add.first)->get_input_port(0), true, 1),
-                                         LoopPort((*sub.first)->get_input_port(0), true, 1)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 1),
-                                         LoopPort((*sub.first)->get_output_port(0), true, 1)});
+        linear_ir_ref->get_loop_manager()->mark_loop(begin_outer, end_outer, 3, 1,
+                                                     std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0), true, 1),
+                                                                           LoopPort((*multiply.first)->get_input_port(1), true, 1),
+                                                                           LoopPort((*add.first)->get_input_port(0), true, 1),
+                                                                           LoopPort((*sub.first)->get_input_port(0), true, 1)},
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 1),
+                                                                           LoopPort((*sub.first)->get_output_port(0), true, 1)});
     }
 }
 
@@ -263,14 +263,14 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsFromInnermostToLoopOuts
         auto add = linear_ir->push_node<ov::opset10::Add>(param_0.second, broadcastmove.second);
         init_expr_descriptors(*add.first, {subtensor, subtensor, subtensor}, {layout, layout, layout});
         auto result = linear_ir->push_node<ov::opset10::Result>(add.second);
-        create_and_add_unified_loop_info(linear_ir, broadcastmove.first, result.first, 3, 1,
-                                        {LoopPort((*broadcastmove.first)->get_input_port(0), true, 1),
-                                         LoopPort((*add.first)->get_input_port(0), true, 1)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 1)});
-        create_and_add_unified_loop_info(linear_ir, broadcastmove.first, result.first, 512, vector_size,
-                                        {LoopPort((*broadcastmove.first)->get_input_port(0), true, 0),
-                                         LoopPort((*add.first)->get_input_port(0), true, 0)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 0)});
+        linear_ir->get_loop_manager()->mark_loop(broadcastmove.first, result.first, 3, 1,
+                                                 std::vector<LoopPort>{LoopPort((*broadcastmove.first)->get_input_port(0), true, 1),
+                                                                       LoopPort((*add.first)->get_input_port(0), true, 1)},
+                                                 std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 1)});
+        linear_ir->get_loop_manager()->mark_loop(broadcastmove.first, result.first, 512, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*broadcastmove.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*add.first)->get_input_port(0), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 0)});
         linear_ir->set_loop_depth(2);
     }
     {
@@ -281,14 +281,14 @@ TEST_F(ExtractLoopInvariantsTest, ExtractedLoopInvariantsFromInnermostToLoopOuts
         auto add = linear_ir_ref->push_node<ov::opset10::Add>(param_0.second, broadcastmove.second);
         init_expr_descriptors(*add.first, {subtensor, subtensor, subtensor}, {layout, layout, layout});
         auto result = linear_ir_ref->push_node<ov::opset10::Result>(add.second);
-        create_and_add_unified_loop_info(linear_ir_ref, add.first, result.first, 3, 1,
-                                        {LoopPort((*add.first)->get_input_port(0), true, 1),
-                                         LoopPort((*add.first)->get_input_port(1), true, 1)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 1)});
-        create_and_add_unified_loop_info(linear_ir_ref, add.first, result.first, 512, vector_size,
-                                        {LoopPort((*add.first)->get_input_port(0), true, 0),
-                                         LoopPort((*add.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*add.first)->get_output_port(0), true, 0)});
+        linear_ir_ref->get_loop_manager()->mark_loop(add.first, result.first, 3, 1,
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_input_port(0), true, 1),
+                                                                           LoopPort((*add.first)->get_input_port(1), true, 1)},
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 1)});
+        linear_ir_ref->get_loop_manager()->mark_loop(add.first, result.first, 512, vector_size,
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_input_port(0), true, 0),
+                                                                           LoopPort((*add.first)->get_input_port(1), true, 0)},
+                                                     std::vector<LoopPort>{LoopPort((*add.first)->get_output_port(0), true, 0)});
     }
 }
 
@@ -356,31 +356,31 @@ TEST_F(ExtractLoopInvariantsRemoveLoopsTest, ExtractedLoopInvariantsAllExprsInLo
         init_expr_descriptors(*multiply.first, {subtensor, subtensor, subtensor}, {layout, layout, layout});
         auto result = linear_ir->push_node<ov::opset10::Result>(multiply.second);
         // 3 inner loop
-        create_and_add_unified_loop_info(linear_ir, max.first, hmax.first, 1, vector_size,
-                                        {LoopPort((*max.first)->get_input_port(0), true, 0),
-                                         LoopPort((*max.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*max.first)->get_output_port(0), true, 0)});
-        create_and_add_unified_loop_info(linear_ir, sub.first, hsum.first, 1, vector_size,
-                                        {LoopPort((*sub.first)->get_input_port(0), true, 0),
-                                         LoopPort((*sub.first)->get_input_port(1), true, 0),
-                                         LoopPort((*add.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*exp.first)->get_output_port(0), true, 0),
-                                         LoopPort((*add.first)->get_output_port(0), true, 0)});
-        create_and_add_unified_loop_info(linear_ir, multiply.first, result.first, 1, vector_size,
-                                        {LoopPort((*multiply.first)->get_input_port(0), true, 0),
-                                         LoopPort((*multiply.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*multiply.first)->get_output_port(0), true, 0)});
+        linear_ir->get_loop_manager()->mark_loop(max.first, hmax.first, 1, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*max.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*max.first)->get_input_port(1), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*max.first)->get_output_port(0), true, 0)});
+        linear_ir->get_loop_manager()->mark_loop(sub.first, hsum.first, 1, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*sub.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*sub.first)->get_input_port(1), true, 0),
+                                                                       LoopPort((*add.first)->get_input_port(1), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*exp.first)->get_output_port(0), true, 0),
+                                                                       LoopPort((*add.first)->get_output_port(0), true, 0)});
+        linear_ir->get_loop_manager()->mark_loop(multiply.first, result.first, 1, vector_size,
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_input_port(0), true, 0),
+                                                                       LoopPort((*multiply.first)->get_input_port(1), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_output_port(0), true, 0)});
         // outer loop info
         const auto loop_begin = std::make_shared<ov::snippets::op::LoopBegin>();
         auto loop_begin_expr = linear_ir->insert_node(loop_begin, std::vector<PortConnectorPtr>{}, {}, false, max.first);
         const auto loop_end = std::make_shared<ov::snippets::op::LoopEnd>();
         std::vector<PortConnectorPtr> loop_end_inputs{(*loop_begin_expr)->get_output_port_connector(0)};
         auto loop_end_expr = linear_ir->insert_node(loop_end, loop_end_inputs, {}, false, result.first);
-        create_and_add_unified_loop_info(linear_ir, loop_begin_expr, result.first, 10, 1,
-                                        {LoopPort((*max.first)->get_input_port(0), true, 1),
-                                         LoopPort((*max.first)->get_input_port(1), true, 0),
-                                         LoopPort((*add.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*multiply.first)->get_output_port(0), true, 1)});
+        linear_ir->get_loop_manager()->mark_loop(loop_begin_expr, result.first, 10, 1,
+                                                 std::vector<LoopPort>{LoopPort((*max.first)->get_input_port(0), true, 1),
+                                                                       LoopPort((*max.first)->get_input_port(1), true, 0),
+                                                                       LoopPort((*add.first)->get_input_port(1), true, 0)},
+                                                 std::vector<LoopPort>{LoopPort((*multiply.first)->get_output_port(0), true, 1)});
         loop_end->set_id((*loop_end_expr)->get_loop_ids().back());
         linear_ir->set_loop_depth(2);
     }
@@ -409,11 +409,11 @@ TEST_F(ExtractLoopInvariantsRemoveLoopsTest, ExtractedLoopInvariantsAllExprsInLo
         const auto loop_end = std::make_shared<ov::snippets::op::LoopEnd>();
         std::vector<PortConnectorPtr> loop_end_inputs{(*loop_begin_expr)->get_output_port_connector(0)};
         auto loop_end_expr = linear_ir_ref->insert_node(loop_end, loop_end_inputs, {}, false, result.first);
-        create_and_add_unified_loop_info(linear_ir_ref, loop_begin_expr, result.first, 10, 1,
-                                        {LoopPort((*max.first)->get_input_port(0), true, 1),
-                                         LoopPort((*max.first)->get_input_port(1), true, 0),
-                                         LoopPort((*add.first)->get_input_port(1), true, 0)},
-                                        {LoopPort((*multiply.first)->get_output_port(0), true, 1)});
+        linear_ir_ref->get_loop_manager()->mark_loop(loop_begin_expr, result.first, 10, 1,
+                                                     std::vector<LoopPort>{LoopPort((*max.first)->get_input_port(0), true, 1),
+                                                                           LoopPort((*max.first)->get_input_port(1), true, 0),
+                                                                           LoopPort((*add.first)->get_input_port(1), true, 0)},
+                                                     std::vector<LoopPort>{LoopPort((*multiply.first)->get_output_port(0), true, 1)});
         loop_end->set_id((*loop_end_expr)->get_loop_ids().back());
     }
 }
diff --git a/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.cpp b/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.cpp
index 1d3cbf022ed299..d062ab1b919ea0 100644
--- a/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.cpp
+++ b/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.cpp
@@ -51,7 +51,7 @@ void CPURuntimeConfigurator::initialization(const std::shared_ptr<ov::snippets::
                             "Incorrect values in subtensors of BrgemmCPU");
 
             if (snippets::utils::is_dynamic_value(*(++in0_subtensor.crbegin())))
-                m_dynamic_brgemms.push_back(expr);
+                m_dynamic_brgemms.insert(expr);
         }
     }
 }
@@ -91,15 +91,8 @@ void CPURuntimeConfigurator::update_brgemms(const ov::snippets::lowered::LoopMan
         const auto& expanded_loop_info = loop_manager->get_loop_info<snippets::lowered::ExpandedLoopInfo>(loop_ids.front());
         const auto& block_size_m = expanded_loop_info->get_work_amount();
 
-        const auto& in_desc = brgemm_expr->get_input_port_descriptor(0);
-        const auto& out_desc = brgemm_expr->get_output_port_descriptor(0);
-
-        auto in_subtensor = in_desc->get_subtensor();
-        auto out_subtensor = out_desc->get_subtensor();
-        *++in_subtensor.rbegin() = block_size_m;
-        *++out_subtensor.rbegin() = block_size_m;
-        in_desc->set_subtensor(in_subtensor);
-        out_desc->set_subtensor(out_subtensor);
+        brgemm_expr->get_input_port_descriptor(0)->set_subtensor_value(1, block_size_m);
+        brgemm_expr->get_output_port_descriptor(0)->set_subtensor_value(1, block_size_m);
     }
 }
 
diff --git a/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.hpp b/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.hpp
index b5ea975bb0c3c0..39ab1977f878d1 100644
--- a/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.hpp
+++ b/src/plugins/intel_cpu/src/emitters/snippets/cpu_runtime_configurator.hpp
@@ -42,16 +42,18 @@ class CPURuntimeConfigurator : public ov::snippets::RuntimeConfigurator {
     void init_tensor_rank(const std::shared_ptr<ov::snippets::lowered::LinearIR>& linear_ir) const override;
     /**
      * @brief Calculate Loop parameters of Loop emitters and update these values in CPURuntimeConfig
-     * @param linear_ir LinearIR
+     * @param loop_manager Loop Manager
      */
     void update_loop_args(const ov::snippets::lowered::LoopManagerPtr& loop_manager) const;
     /**
      * @brief Update subtensors of Brgemms
+     * @param loop_manager Loop Manager
      */
     void update_brgemms(const ov::snippets::lowered::LoopManagerPtr& loop_manager) const;
 
     const size_t rank6D = 6;
-    std::vector<ov::snippets::lowered::ExpressionPtr> m_dynamic_brgemms = {};
+    // Brgemm expressions with subtensors with dynamic values
+    std::unordered_set<ov::snippets::lowered::ExpressionPtr> m_dynamic_brgemms = {};
 };
 
 }   // namespace intel_cpu
diff --git a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.cpp b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.cpp
index 5e4cb94d5bdd4f..03f63f2fb6e603 100644
--- a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.cpp
+++ b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.cpp
@@ -57,6 +57,15 @@ LinearIR::constExprIt BrgemmBlocking::get_loop_begin_pos(LinearIR& linear_ir, co
     return loop_begin_it;
 }
 
+snippets::lowered::SpecificIterationHandlers BrgemmBlocking::get_default_blocking_loop_handlers(size_t work_amount, size_t block_size) {
+    SpecificIterationHandlers handlers;
+    const auto tail_size = snippets::utils::is_dynamic_value(work_amount) ? snippets::utils::get_dynamic_value<size_t>() : work_amount % block_size;
+    if (tail_size != 0)
+        handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, snippets::lowered::pass::UpdateSubtensors>(tail_size);
+    handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, SetEvaluanceOnce>();
+    return handlers;
+}
+
 bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, LinearIR::constExprIt end) {
     OV_ITT_SCOPED_TASK(ov::pass::itt::domains::SnippetsTransform, "Snippets::BrgemmBlocking")
     const auto& loop_manager = linear_ir.get_loop_manager();
@@ -111,22 +120,18 @@ bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, Linea
         // If block_size is dynamic, it means that Brgemm will process full tensor:
         //   subtensor[i] = FULL_DIM as by default
         if (!snippets::utils::is_dynamic_value(block_size_m)) {
-            *++in_0_subtensor.rbegin() = block_size_m;
-            *++out_subtensor.rbegin() = block_size_m;
+            brgemm_expr->get_input_port_descriptor(0)->set_subtensor_value(1, block_size_m);
+            brgemm_expr->get_output_port_descriptor(0)->set_subtensor_value(1, block_size_m);
         }
         if (!snippets::utils::is_dynamic_value(block_size_n)) {
-            *in_1_subtensor.rbegin() = block_size_n;
-            *out_subtensor.rbegin() = block_size_n;
+            brgemm_expr->get_input_port_descriptor(1)->set_subtensor_value(0, block_size_n);
+            brgemm_expr->get_output_port_descriptor(0)->set_subtensor_value(0, block_size_n);
         }
         if (!snippets::utils::is_dynamic_value(block_size_k)) {
-            *in_0_subtensor.rbegin() = block_size_k;
-            *++in_1_subtensor.rbegin() = block_size_k;
+            brgemm_expr->get_input_port_descriptor(0)->set_subtensor_value(0, block_size_k);
+            brgemm_expr->get_input_port_descriptor(1)->set_subtensor_value(1, block_size_k);
         }
 
-        brgemm_expr->get_input_port_descriptor(0)->set_subtensor(in_0_subtensor);
-        brgemm_expr->get_input_port_descriptor(1)->set_subtensor(in_1_subtensor);
-        brgemm_expr->get_output_port_descriptor(0)->set_subtensor(out_subtensor);
-
         ov::snippets::lowered::ExpressionPtr copy_b_expr = nullptr;
         if (brgemm_cpu && brgemm_cpu->is_with_data_repacking()) {
             const auto copy_b = brgemm_cpu->get_brgemm_copy();
@@ -150,15 +155,6 @@ bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, Linea
             }
         }
 
-        auto get_default_handlers = [](size_t work_amount, size_t block_size) {
-            SpecificIterationHandlers handlers;
-            const auto tail_size = snippets::utils::is_dynamic_value(work_amount) ? snippets::utils::get_dynamic_value<size_t>() : work_amount % block_size;
-            if (tail_size != 0)
-                handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, snippets::lowered::pass::UpdateSubtensors>(tail_size);
-            handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, SetEvaluanceOnce>(true);
-            return handlers;
-        };
-
         auto mark_m_blocking = [&](bool include_repacking) {
             const auto loop_begin_it = get_loop_begin_pos(linear_ir, expr_it, include_repacking);
             const auto loop_end_it = std::next(expr_it);
@@ -173,7 +169,7 @@ bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, Linea
             const std::vector<LoopPort> exits{LoopPort(brgemm_expr->get_output_port(0), true)};
 
             const auto id = loop_manager->mark_loop(loop_begin_it, loop_end_it, m, block_size_m, 1, entries, exits, false);
-            loop_manager->get_loop_info<ov::snippets::lowered::UnifiedLoopInfo>(id)->set_handlers(get_default_handlers(m, block_size_m));
+            loop_manager->get_loop_info<ov::snippets::lowered::UnifiedLoopInfo>(id)->set_handlers(get_default_blocking_loop_handlers(m, block_size_m));
         };
 
         auto mark_n_blocking = [&]() {
@@ -186,7 +182,7 @@ bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, Linea
             const std::vector<LoopPort> exits{LoopPort(brgemm_expr->get_output_port(0), true)};
 
             const auto id = loop_manager->mark_loop(loop_begin_it, loop_end_it, n, block_size_n, 0, entries, exits, false);
-            loop_manager->get_loop_info<ov::snippets::lowered::UnifiedLoopInfo>(id)->set_handlers(get_default_handlers(n, block_size_n));
+            loop_manager->get_loop_info<ov::snippets::lowered::UnifiedLoopInfo>(id)->set_handlers(get_default_blocking_loop_handlers(n, block_size_n));
         };
 
         auto mark_k_blocking = [&]() {
@@ -198,7 +194,7 @@ bool BrgemmBlocking::run(LinearIR& linear_ir, LinearIR::constExprIt begin, Linea
                 LoopPort(brgemm_cpu && brgemm_cpu->is_with_data_repacking() ? copy_b_expr->get_input_port(0) : brgemm_expr->get_input_port(1), true, 1)};
             const std::vector<LoopPort> exits{LoopPort(brgemm_expr->get_output_port(0), false)};
 
-            auto handlers = get_default_handlers(k, block_size_k);
+            auto handlers = get_default_blocking_loop_handlers(k, block_size_k);
             handlers.register_pass<ov::snippets::lowered::SpecificLoopIterType::FIRST_ITER, SetBrgemmBeta>(0.f);
 
             const auto id = loop_manager->mark_loop(loop_begin_it, loop_end_it, k, block_size_k, entries, exits, false);
diff --git a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.hpp b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.hpp
index cdc2d05cffd1e5..4d29267f034fc9 100644
--- a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.hpp
+++ b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/brgemm_blocking.hpp
@@ -5,6 +5,7 @@
 #pragma once
 
 #include "snippets/lowered/pass/pass.hpp"
+#include "snippets/lowered/specific_loop_iter_handlers.hpp"
 
 namespace ov {
 namespace intel_cpu {
@@ -24,6 +25,8 @@ class BrgemmBlocking : public snippets::lowered::pass::RangedPass {
              snippets::lowered::LinearIR::constExprIt begin,
              snippets::lowered::LinearIR::constExprIt end) override;
 
+    static snippets::lowered::SpecificIterationHandlers get_default_blocking_loop_handlers(size_t work_amount, size_t block_size);
+
 private:
     static snippets::lowered::LinearIR::constExprIt move_new_memory_buffer(snippets::lowered::LinearIR& linear_ir,
                                                                            const snippets::lowered::LinearIR::constExprIt& brgemm_it);
diff --git a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.cpp b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.cpp
index 97c8c2a5299f6b..6acf68fe245467 100644
--- a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.cpp
+++ b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.cpp
@@ -36,24 +36,16 @@ std::shared_ptr<snippets::lowered::pass::PassBase> SetBrgemmBeta::merge(const st
     return merged_pass;
 }
 
-SetEvaluanceOnce::SetEvaluanceOnce(bool evaluation) : snippets::lowered::pass::RangedPass(), m_evaluation(evaluation) {}
-
 bool SetEvaluanceOnce::run(LinearIR& linear_ir, LinearIR::constExprIt begin, LinearIR::constExprIt end) {
     const auto& loop_end = ov::as_type_ptr<snippets::op::LoopEnd>(end->get()->get_node());
     OPENVINO_ASSERT(loop_end, "SetEvaluanceOnce expected LoopEnd node in iterator `end`.");
     const auto& loop_info = linear_ir.get_loop_manager()->get_loop_info<ov::snippets::lowered::ExpandedLoopInfo>(loop_end->get_id());
-    loop_info->set_evaluate_once(m_evaluation);
+    loop_info->set_evaluate_once(true);
     return true;
 }
 
 std::shared_ptr<snippets::lowered::pass::PassBase> SetEvaluanceOnce::merge(const std::shared_ptr<snippets::lowered::pass::PassBase>& other) {
-    const auto merged_pass = std::make_shared<SetEvaluanceOnce>(m_evaluation);
-    if (other == nullptr)
-        return merged_pass;
-    const auto casted_pass = ov::as_type_ptr<SetEvaluanceOnce>(other);
-    if (!casted_pass || m_evaluation != casted_pass->m_evaluation)
-        return nullptr;
-    return merged_pass;
+    return !other || ov::is_type<SetEvaluanceOnce>(other) ? std::make_shared<SetEvaluanceOnce>() : nullptr;
 }
 
 }  // namespace pass
diff --git a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.hpp b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.hpp
index 7616954bc2cca5..2a9492713e0c46 100644
--- a/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.hpp
+++ b/src/plugins/intel_cpu/src/transformations/snippets/x64/pass/lowered/cpu_iter_handlers.hpp
@@ -30,22 +30,18 @@ class SetBrgemmBeta : public snippets::lowered::pass::RangedPass {
 
 /**
  * @interface SetEvaluanceOnce
- * @brief The pass set `evaluate once` only to ExpandedLoopInfo which is mapped on LoopEnd in the passed iterator `end`.
+ * @brief The pass set `evaluate once = true` only to ExpandedLoopInfo which is mapped on LoopEnd in the passed iterator `end`.
  *        The pointer arithmetic should be updated in the separate optimization `OptimizeLoopSingleEvaluation`
- * @param m_evaluation - value which must be set
  * @ingroup snippets
  */
 class SetEvaluanceOnce : public snippets::lowered::pass::RangedPass {
 public:
-    SetEvaluanceOnce(bool evaluation);
+    SetEvaluanceOnce() = default;
     OPENVINO_RTTI("SetEvaluanceOnce", "RangedPass")
     bool run(snippets::lowered::LinearIR& linear_ir,
              snippets::lowered::LinearIR::constExprIt begin,
              snippets::lowered::LinearIR::constExprIt end) override;
     std::shared_ptr<snippets::lowered::pass::PassBase> merge(const std::shared_ptr<snippets::lowered::pass::PassBase>& other) override;
-
-private:
-    bool m_evaluation = false;
 };
 }  // namespace pass
 }  // namespace intel_cpu
diff --git a/src/plugins/intel_cpu/tests/functional/shared_tests_instances/snippets/matmul.cpp b/src/plugins/intel_cpu/tests/functional/shared_tests_instances/snippets/matmul.cpp
index 8e1959224ef6a1..1089bdc3faffaa 100644
--- a/src/plugins/intel_cpu/tests/functional/shared_tests_instances/snippets/matmul.cpp
+++ b/src/plugins/intel_cpu/tests/functional/shared_tests_instances/snippets/matmul.cpp
@@ -96,7 +96,7 @@ std::vector<std::vector<ov::test::InputShape>> input_shapes_dynamic{
         },
         // Only N dimension is dynamic
         {
-            STATIC_SHAPE(2, 2, 65, 550),
+            {PartialShape{},              {{2, 2, 65, 550}}},
             {PartialShape{2, 2, 550, -1}, {{2, 2, 550, 70}, {2, 2, 550, 12}, {2, 2, 550, 70},
                                            {2, 2, 550, 12}, {2, 2, 550, 10}}}
         },
diff --git a/src/plugins/intel_cpu/tests/unit/snippets_transformations/x64/lowered/brgemm_blocking.cpp b/src/plugins/intel_cpu/tests/unit/snippets_transformations/x64/lowered/brgemm_blocking.cpp
index 9b55564ea09186..f416d06791e182 100644
--- a/src/plugins/intel_cpu/tests/unit/snippets_transformations/x64/lowered/brgemm_blocking.cpp
+++ b/src/plugins/intel_cpu/tests/unit/snippets_transformations/x64/lowered/brgemm_blocking.cpp
@@ -7,8 +7,8 @@
 #include "lir_test_utils.hpp"
 #include "openvino/opsets/opset10.hpp"
 #include "snippets/lowered/linear_ir.hpp"
+#include "snippets/lowered/loop_info.hpp"
 #include "snippets/lowered/pass/propagate_subtensors.hpp"
-#include "snippets/lowered/pass/serialize_control_flow.hpp"
 #include "snippets/snippets_isa.hpp"
 #include "transformations/snippets/x64/op/brgemm_copy_b.hpp"
 #include "transformations/snippets/x64/op/brgemm_cpu.hpp"
@@ -23,14 +23,6 @@ using namespace ov::snippets::lowered;
 using namespace ov::snippets;
 
 namespace {
-SpecificIterationHandlers get_default_handlers(size_t work_amount, size_t block_size) {
-    SpecificIterationHandlers handlers;
-    const auto tail_size = snippets::utils::is_dynamic_value(work_amount) ? snippets::utils::get_dynamic_value<size_t>() : work_amount % block_size;
-    if (tail_size != 0)
-        handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, snippets::lowered::pass::UpdateSubtensors>(tail_size);
-    handlers.register_pass<snippets::lowered::SpecificLoopIterType::LAST_ITER, ov::intel_cpu::pass::SetEvaluanceOnce>(true);
-    return handlers;
-}
 
 void create_brgemm_loop_infos(const LinearIRPtr& linear_ir,
                               const ExpressionPtr& brgemm_expr,
@@ -41,29 +33,30 @@ void create_brgemm_loop_infos(const LinearIRPtr& linear_ir,
     const bool n_block = k != 0 && k_blk != 0;
     const bool m_block = m != 0 && m_blk != 0;
     if (k_block) {
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, k, k_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0)),
-                                                                 LoopPort(brgemm_expr->get_input_port(1), true, 1)},
-                                                                {LoopPort(brgemm_expr->get_output_port(0), false)}, false);
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(k, k_block));
+        const auto loop_info =
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(k, k_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0)),
+                                      LoopPort(brgemm_expr->get_input_port(1), true, 1)},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0), false)},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(k, k_block));
         loop_info->register_pass_to_handler<SpecificLoopIterType::FIRST_ITER, ov::intel_cpu::pass::SetBrgemmBeta>(0.f);
+        linear_ir->get_loop_manager()->add_loop_info(loop_info);
     }
     if (n_block) {
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, n, n_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0), false),
-                                                                 LoopPort(brgemm_expr->get_input_port(1))},
-                                                                {LoopPort(brgemm_expr->get_output_port(0))}, false);
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(n, n_block));
+        linear_ir->get_loop_manager()->add_loop_info(
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(n, n_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0), false),
+                                      LoopPort(brgemm_expr->get_input_port(1))},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0))},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(n, n_block)));
     }
     if (m_block) {
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, m, m_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0), true, 1),
-                                                                 LoopPort(brgemm_expr->get_input_port(1), false, 1)},
-                                                                {LoopPort(brgemm_expr->get_output_port(0), true, 1)}, false);
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(m, m_block));
+        linear_ir->get_loop_manager()->add_loop_info(
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(m, m_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0), true, 1),
+                                      LoopPort(brgemm_expr->get_input_port(1), false, 1)},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0), true, 1)},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(m, m_block)));
     }
 }
 
@@ -77,30 +70,31 @@ void create_brgemm_with_copy_b_loop_infos(const LinearIRPtr& linear_ir,
     const bool n_block = k != 0 && k_blk != 0;
     const bool m_block = m != 0 && m_blk != 0;
     if (k_block) {
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, k, k_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0)),
-                                                                 LoopPort(copy_b_expr->get_input_port(0), true, 1)},
-                                                                {LoopPort(brgemm_expr->get_output_port(0), false)});
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(k, k_block));
+        const auto loop_info =
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(k, k_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0)),
+                                      LoopPort(copy_b_expr->get_input_port(0), true, 1)},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0), false)},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(k, k_block));
         loop_info->register_pass_to_handler<SpecificLoopIterType::FIRST_ITER, ov::intel_cpu::pass::SetBrgemmBeta>(0.f);
+        linear_ir->get_loop_manager()->add_loop_info(loop_info);
     }
     if (n_block) {
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, n, n_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0), false),
-                                                                 LoopPort(copy_b_expr->get_input_port(0))},
-                                                                {LoopPort(brgemm_expr->get_output_port(0))});
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(n, n_block));
+        linear_ir->get_loop_manager()->add_loop_info(
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(n, n_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0), false),
+                                      LoopPort(copy_b_expr->get_input_port(0))},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0))},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(n, n_block)));
     }
     if (m_block) {
         const auto& second_input_port = k_block || n_block ? copy_b_expr->get_input_port(0) : brgemm_expr->get_input_port(1);
-        const size_t loop_id = create_and_add_unified_loop_info(linear_ir, m, m_blk,
-                                                                {LoopPort(brgemm_expr->get_input_port(0), true, 1),
-                                                                 LoopPort(second_input_port, false, 1)},
-                                                                {LoopPort(brgemm_expr->get_output_port(0), true, 1)});
-        const auto& loop_info = linear_ir->get_loop_manager()->get_loop_info<UnifiedLoopInfo>(loop_id);
-        loop_info->set_handlers(get_default_handlers(m, m_block));
+        linear_ir->get_loop_manager()->add_loop_info(
+            std::make_shared<ov::snippets::lowered::UnifiedLoopInfo>(m, m_blk,
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_input_port(0), true, 1),
+                                      LoopPort(second_input_port, false, 1)},
+                std::vector<LoopPort>{LoopPort(brgemm_expr->get_output_port(0), true, 1)},
+                ov::intel_cpu::pass::BrgemmBlocking::get_default_blocking_loop_handlers(m, m_block)));
     }
 }
 } // namespace