diff --git a/bigframes/core/compile/scalar_op_compiler.py b/bigframes/core/compile/scalar_op_compiler.py
index a52264be17..4d43545efe 100644
--- a/bigframes/core/compile/scalar_op_compiler.py
+++ b/bigframes/core/compile/scalar_op_compiler.py
@@ -622,6 +622,26 @@ def strftime_op_impl(x: ibis_types.Value, op: ops.StrftimeOp):
     )
 
 
+@scalar_op_compiler.register_unary_op(ops.FloorDtOp, pass_op=True)
+def floor_dt_op_impl(x: ibis_types.Value, op: ops.FloorDtOp):
+    supported_freqs = ["Y", "Q", "M", "W", "D", "h", "min", "s", "ms", "us", "ns"]
+    pandas_to_ibis_freqs = {"min": "m"}
+    if op.freq not in supported_freqs:
+        raise NotImplementedError(
+            f"Unsupported freq paramater: {op.freq}"
+            + " Supported freq parameters are: "
+            + ",".join(supported_freqs)
+        )
+    if op.freq in pandas_to_ibis_freqs:
+        ibis_freq = pandas_to_ibis_freqs[op.freq]
+    else:
+        ibis_freq = op.freq
+    result_type = x.type()
+    result = typing.cast(ibis_types.TimestampValue, x)
+    result = result.truncate(ibis_freq)
+    return result.cast(result_type)
+
+
 @scalar_op_compiler.register_unary_op(ops.time_op)
 def time_op_impl(x: ibis_types.Value):
     return typing.cast(ibis_types.TimestampValue, x).time()
diff --git a/bigframes/operations/__init__.py b/bigframes/operations/__init__.py
index dbea6145e0..f2bcbd894a 100644
--- a/bigframes/operations/__init__.py
+++ b/bigframes/operations/__init__.py
@@ -441,6 +441,15 @@ def output_type(self, *input_types):
         return dtypes.STRING_DTYPE
 
 
+@dataclasses.dataclass(frozen=True)
+class FloorDtOp(UnaryOp):
+    name: typing.ClassVar[str] = "floor_dt"
+    freq: str
+
+    def output_type(self, *input_types):
+        return input_types[0]
+
+
 # Binary Ops
 fillna_op = create_binary_op(name="fillna")
 cliplower_op = create_binary_op(name="clip_lower")
diff --git a/bigframes/operations/datetimes.py b/bigframes/operations/datetimes.py
index 1b4a2fe0e6..7d25ac3622 100644
--- a/bigframes/operations/datetimes.py
+++ b/bigframes/operations/datetimes.py
@@ -97,3 +97,6 @@ def strftime(self, date_format: str) -> series.Series:
 
     def normalize(self) -> series.Series:
         return self._apply_unary_op(ops.normalize_op)
+
+    def floor(self, freq: str) -> series.Series:
+        return self._apply_unary_op(ops.FloorDtOp(freq=freq))
diff --git a/tests/system/small/operations/test_datetimes.py b/tests/system/small/operations/test_datetimes.py
index d5100e7dc2..b952289a72 100644
--- a/tests/system/small/operations/test_datetimes.py
+++ b/tests/system/small/operations/test_datetimes.py
@@ -282,3 +282,24 @@ def test_dt_normalize(scalars_dfs, col_name):
         pd_result.astype(scalars_df[col_name].dtype),  # normalize preserves type
         bf_result,
     )
+
+
+@pytest.mark.parametrize(
+    ("col_name", "freq"),
+    [
+        ("timestamp_col", "D"),
+        ("timestamp_col", "min"),
+        ("datetime_col", "s"),
+        ("datetime_col", "us"),
+    ],
+)
+@skip_legacy_pandas
+def test_dt_floor(scalars_dfs, col_name, freq):
+    scalars_df, scalars_pandas_df = scalars_dfs
+    bf_result = scalars_df[col_name].dt.floor(freq).to_pandas()
+    pd_result = scalars_pandas_df[col_name].dt.floor(freq)
+
+    assert_series_equal(
+        pd_result.astype(scalars_df[col_name].dtype),  # floor preserves type
+        bf_result,
+    )
diff --git a/third_party/bigframes_vendored/pandas/core/arrays/datetimelike.py b/third_party/bigframes_vendored/pandas/core/arrays/datetimelike.py
index 60ac19b818..f706ae2560 100644
--- a/third_party/bigframes_vendored/pandas/core/arrays/datetimelike.py
+++ b/third_party/bigframes_vendored/pandas/core/arrays/datetimelike.py
@@ -67,3 +67,33 @@ def normalize(self):
             bigframes.series.Series of the same dtype as the data.
         """
         raise NotImplementedError(constants.ABSTRACT_METHOD_ERROR_MESSAGE)
+
+    def floor(self, freq: str):
+        """
+        Perform floor operation on the data to the specified freq.
+
+        Supported freq arguments are: 'Y' (year), 'Q' (quarter), 'M'
+        (month), 'W' (week), 'D' (day), 'h' (hour), 'min' (minute), 's'
+        (second), 'ms' (microsecond), 'us' (nanosecond), 'ns' (nanosecond)
+
+        Behavior around clock changes (i.e. daylight savings) is determined
+        by the SQL engine, so "ambiguous" and "nonexistent" parameters are not
+        supported. Y, Q, M, and W freqs are not supported by pandas as of
+        version 2.2, but have been added here due to backend support.
+
+        **Examples:**
+
+            >>> import pandas as pd
+            >>> import bigframes.pandas as bpd
+            >>> rng = pd.date_range('1/1/2018 11:59:00', periods=3, freq='min')
+            >>> bpd.Series(rng).dt.floor("h")
+            0    2018-01-01 11:00:00
+            1    2018-01-01 12:00:00
+            2    2018-01-01 12:00:00
+            dtype: timestamp[us][pyarrow]
+
+        Args:
+            freq (str):
+                Frequency string (e.g. "D", "min", "s").
+        """
+        raise NotImplementedError(constants.ABSTRACT_METHOD_ERROR_MESSAGE)