diff --git a/src/plugins/intel_gpu/src/graph/graph_optimizer/compile_graph.cpp b/src/plugins/intel_gpu/src/graph/graph_optimizer/compile_graph.cpp
index c749edc9784de4..5359ca294a3b30 100644
--- a/src/plugins/intel_gpu/src/graph/graph_optimizer/compile_graph.cpp
+++ b/src/plugins/intel_gpu/src/graph/graph_optimizer/compile_graph.cpp
@@ -20,9 +20,8 @@ using namespace cldnn;
 
 void compile_graph::run(program& p) {
     OV_ITT_SCOPED_TASK(itt::domains::CLDNN, "CLDNN::pass::CompileGraph");
-    size_t order_idx = 0;
     for (auto& node : p.get_processing_order()) {
-        node->set_unique_id(std::to_string(order_idx++));
+        node->set_unique_id();
         if (!node->is_type<data>()) {
             node->get_output_layout();
         }
diff --git a/src/plugins/intel_gpu/src/graph/graph_optimizer/post_input_reorder.cpp b/src/plugins/intel_gpu/src/graph/graph_optimizer/post_input_reorder.cpp
index a8ff3c4d739899..d2ab615804db38 100644
--- a/src/plugins/intel_gpu/src/graph/graph_optimizer/post_input_reorder.cpp
+++ b/src/plugins/intel_gpu/src/graph/graph_optimizer/post_input_reorder.cpp
@@ -62,7 +62,7 @@ void post_input_reorder::run(program& p) {
                                       input_layout.size,
                                       input_layout.data_padding);
                 auto& reorder = add_reorder(p, input, node, current_layout);
-                reorder.set_unique_id(node->get_unique_id() + "_input_reorder");
+                reorder.set_unique_id();
                 reorder.get_output_layout(false);
                 node->set_output_layout(previous_layout, false);
                 reorder.set_selected_impl(reorder.type()->choose_impl(reorder));
diff --git a/src/plugins/intel_gpu/src/graph/graph_optimizer/remove_redundant_reorders.cpp b/src/plugins/intel_gpu/src/graph/graph_optimizer/remove_redundant_reorders.cpp
index eb6b8ca1ea6742..a4476bf6b17bf3 100644
--- a/src/plugins/intel_gpu/src/graph/graph_optimizer/remove_redundant_reorders.cpp
+++ b/src/plugins/intel_gpu/src/graph/graph_optimizer/remove_redundant_reorders.cpp
@@ -35,7 +35,7 @@ void remove_redundant_reorders::run(program& p) {
         if (!update_implementations)
             return;
 
-        node.set_unique_id(node.get_unique_id() + "_reorder");
+        node.set_unique_id();
         auto new_impl = node.type()->choose_impl(node);
         node.set_selected_impl(std::move(new_impl));
     };
diff --git a/src/plugins/intel_gpu/src/graph/include/program_node.h b/src/plugins/intel_gpu/src/graph/include/program_node.h
index 6c80fc6958aae5..bc0d0c0607f84e 100644
--- a/src/plugins/intel_gpu/src/graph/include/program_node.h
+++ b/src/plugins/intel_gpu/src/graph/include/program_node.h
@@ -7,6 +7,7 @@
 #include "intel_gpu/primitives/primitive.hpp"
 #include "intel_gpu/primitives/activation.hpp"
 #include "intel_gpu/primitives/implementation_desc.hpp"
+#include "intel_gpu/graph/program.hpp"
 
 #include "kernel_selector_helper.h"
 #include "meta_utils.h"
@@ -17,6 +18,7 @@
 #include <memory>
 #include <list>
 #include <algorithm>
+#include <thread>
 
 namespace cldnn {
 
@@ -350,11 +352,19 @@ struct program_node {
 
     bool need_lockable_memory() const;
 
-    std::string get_unique_id() const { return unique_id; }
-    void set_unique_id(std::string id) { unique_id = id; }
+    size_t get_unique_id() const { return unique_id; }
+
+    void set_unique_id() {
+        unique_id = cur_id++;
+    }
+
+    static void reset_unique_id() {
+        cur_id = 0;
+    }
 
 protected:
-    std::string unique_id;
+    size_t unique_id = 0;
+    static thread_local size_t cur_id;
 
     std::shared_ptr<primitive> desc;
     program& myprog;
diff --git a/src/plugins/intel_gpu/src/graph/kernel_selector_helper.cpp b/src/plugins/intel_gpu/src/graph/kernel_selector_helper.cpp
index cab06f3e87ed62..66bc73c0afee77 100644
--- a/src/plugins/intel_gpu/src/graph/kernel_selector_helper.cpp
+++ b/src/plugins/intel_gpu/src/graph/kernel_selector_helper.cpp
@@ -839,7 +839,7 @@ void set_params(const program_node& node, kernel_selector::params& params) {
     const auto& program = node.get_program();
     const auto& device_info = program.get_engine().get_device_info();
 
-    params.uniqueID = std::to_string(program.get_id()) + "_"  + node.get_unique_id();
+    params.uniqueID = std::to_string(node.get_unique_id());
     params.engineInfo.bSubGroupSupport = device_info.supports_subgroups;
     params.engineInfo.bSubGroupShortSupport = device_info.supports_subgroups_short;
     params.engineInfo.bSubGroupCharSupport = device_info.supports_subgroups_char;
diff --git a/src/plugins/intel_gpu/src/graph/program.cpp b/src/plugins/intel_gpu/src/graph/program.cpp
index 2518bd8903a780..73fcf10162e70a 100644
--- a/src/plugins/intel_gpu/src/graph/program.cpp
+++ b/src/plugins/intel_gpu/src/graph/program.cpp
@@ -107,6 +107,7 @@ program::program(engine& engine_ref,
     prepare_nodes(topology);
     _kernels_cache = std::unique_ptr<kernels_cache>(new kernels_cache(_engine, prog_id,
                                                                       kernel_selector::KernelBase::get_db().get_batch_header_str()));
+    program_node::reset_unique_id();
     if (no_optimizations) {
         init_graph();
     } else {
diff --git a/src/plugins/intel_gpu/src/graph/program_node.cpp b/src/plugins/intel_gpu/src/graph/program_node.cpp
index 61861e42bee8c5..5d1dc6a6aba01f 100644
--- a/src/plugins/intel_gpu/src/graph/program_node.cpp
+++ b/src/plugins/intel_gpu/src/graph/program_node.cpp
@@ -3,7 +3,6 @@
 //
 
 #include "program_node.h"
-#include "intel_gpu/graph/program.hpp"
 #include "program_helpers.h"
 #include "primitive_inst.h"
 
@@ -24,6 +23,8 @@
 
 using namespace cldnn;
 
+thread_local size_t program_node::cur_id = 0;
+
 program_node::program_node(std::shared_ptr<primitive> prim, program& prog)
     : desc(prim), myprog(prog), org_id(prim ? (prim->id) : 0) {
     if (prim)
diff --git a/src/tests/functional/plugin/gpu/behavior/cache.cpp b/src/tests/functional/plugin/gpu/behavior/cache.cpp
index 2d64e8812b9ba1..853e2612497e39 100644
--- a/src/tests/functional/plugin/gpu/behavior/cache.cpp
+++ b/src/tests/functional/plugin/gpu/behavior/cache.cpp
@@ -49,6 +49,39 @@ TEST_F(CompiledKernelsCacheTest, CanCreateCacheDirAndDumpBinaries) {
     }
 }
 
+TEST_F(CompiledKernelsCacheTest, TwoNetworksWithSameModelCreatesSameCache) {
+    std::shared_ptr<InferenceEngine::Core> ie = PluginCache::get().ie();
+    // Create two CNNNetwork from same ngraph::Function
+    InferenceEngine::CNNNetwork cnnNet1(function);
+    InferenceEngine::CNNNetwork cnnNet2(function);
+    std::map<std::string, std::string> config = {{ CONFIG_KEY(CACHE_DIR), cache_path }};
+    try {
+        // Load 1st CNNNetwork
+        auto execNet1 = ie->LoadNetwork(cnnNet1, "GPU", config);
+        auto n_cache_files = CommonTestUtils::listFilesWithExt(cache_path, "cl_cache").size();
+
+        // Check that directory with cached kernels exists after loading network
+        ASSERT_TRUE(CommonTestUtils::directoryExists(cache_path)) << "Directory with cached kernels doesn't exist";
+        // Load 2nd CNNNetwork
+        auto execNet2 = ie->LoadNetwork(cnnNet2, "GPU", config);
+
+        // Check that two loaded networks with same function creates same caches
+        ASSERT_EQ(CommonTestUtils::removeFilesWithExt(cache_path, "cl_cache"), n_cache_files);
+
+        // Remove directory and check that it doesn't exist anymore
+        ASSERT_EQ(CommonTestUtils::removeDir(cache_path), 0);
+        ASSERT_FALSE(CommonTestUtils::directoryExists(cache_path));
+    } catch (std::exception& ex) {
+        // Cleanup in case of any exception
+        if (CommonTestUtils::directoryExists(cache_path)) {
+            ASSERT_GE(CommonTestUtils::removeFilesWithExt(cache_path, "cl_cache"), 0);
+            ASSERT_EQ(CommonTestUtils::removeDir(cache_path), 0);
+        }
+        FAIL() << ex.what() << std::endl;
+    }
+}
+
+
 #ifdef OPENVINO_ENABLE_UNICODE_PATH_SUPPORT
 
 TEST_F(CompiledKernelsCacheTest, CanCreateCacheDirAndDumpBinariesUnicodePath) {