diff --git a/bigframes/ml/decomposition.py b/bigframes/ml/decomposition.py
index ece950a5a2..3ff32d2433 100644
--- a/bigframes/ml/decomposition.py
+++ b/bigframes/ml/decomposition.py
@@ -360,5 +360,12 @@ def score(
         if not self._bqml_model:
             raise RuntimeError("A model must be fitted before score")
 
-        # TODO(b/291973741): X param is ignored. Update BQML supports input in ML.EVALUATE.
-        return self._bqml_model.evaluate()
+        if X is not None and y is not None:
+            X, y = utils.batch_convert_to_dataframe(
+                X, y, session=self._bqml_model.session
+            )
+            input_data = X.join(y, how="outer")
+        else:
+            input_data = X
+
+        return self._bqml_model.evaluate(input_data)
diff --git a/tests/system/large/ml/test_decomposition.py b/tests/system/large/ml/test_decomposition.py
index d1a5f9f2aa..e0e4b79c6f 100644
--- a/tests/system/large/ml/test_decomposition.py
+++ b/tests/system/large/ml/test_decomposition.py
@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import pandas as pd
+import pandas.testing
 
 from bigframes.ml import decomposition
 from tests.system import utils
@@ -193,7 +194,16 @@ def test_decomposition_mf_configure_fit_load(
         )
     )
 
-    reloaded_model.score(new_ratings)
+    # Make sure the input to score is not ignored.
+    scores_training_data = reloaded_model.score().to_pandas()
+    scores_new_ratings = reloaded_model.score(new_ratings).to_pandas()
+    pandas.testing.assert_index_equal(
+        scores_training_data.columns, scores_new_ratings.columns
+    )
+    assert (
+        scores_training_data["mean_squared_error"].iloc[0]
+        != scores_new_ratings["mean_squared_error"].iloc[0]
+    )
 
     result = reloaded_model.predict(new_ratings).to_pandas()
 
diff --git a/tests/unit/ml/test_golden_sql.py b/tests/unit/ml/test_golden_sql.py
index 62cfe09704..10fefcc457 100644
--- a/tests/unit/ml/test_golden_sql.py
+++ b/tests/unit/ml/test_golden_sql.py
@@ -81,6 +81,7 @@ def mock_X(mock_y, mock_session):
         ["index_column_id"],
         ["index_column_label"],
     )
+    type(mock_X).sql = mock.PropertyMock(return_value="input_X_sql_property")
     mock_X.reset_index(drop=True).cache().sql = "input_X_no_index_sql"
     mock_X.join(mock_y).sql = "input_X_y_sql"
     mock_X.join(mock_y).cache.return_value = mock_X.join(mock_y)
@@ -248,7 +249,7 @@ def test_decomposition_mf_predict(mock_session, bqml_model, mock_X):
     )
 
 
-def test_decomposition_mf_score(mock_session, bqml_model, mock_X):
+def test_decomposition_mf_score(mock_session, bqml_model):
     model = decomposition.MatrixFactorization(
         num_factors=34,
         feedback_type="explicit",
@@ -258,8 +259,23 @@ def test_decomposition_mf_score(mock_session, bqml_model, mock_X):
         l2_reg=9.83,
     )
     model._bqml_model = bqml_model
-    model.score(mock_X)
-
+    model.score()
     mock_session.read_gbq.assert_called_once_with(
         "SELECT * FROM ML.EVALUATE(MODEL `model_project`.`model_dataset`.`model_id`)"
     )
+
+
+def test_decomposition_mf_score_with_x(mock_session, bqml_model, mock_X):
+    model = decomposition.MatrixFactorization(
+        num_factors=34,
+        feedback_type="explicit",
+        user_col="user_id",
+        item_col="item_col",
+        rating_col="rating_col",
+        l2_reg=9.83,
+    )
+    model._bqml_model = bqml_model
+    model.score(mock_X)
+    mock_session.read_gbq.assert_called_once_with(
+        "SELECT * FROM ML.EVALUATE(MODEL `model_project`.`model_dataset`.`model_id`,\n  (input_X_sql_property))"
+    )
diff --git a/third_party/bigframes_vendored/sklearn/decomposition/_mf.py b/third_party/bigframes_vendored/sklearn/decomposition/_mf.py
index fb29cc8984..c3c3a77b71 100644
--- a/third_party/bigframes_vendored/sklearn/decomposition/_mf.py
+++ b/third_party/bigframes_vendored/sklearn/decomposition/_mf.py
@@ -73,11 +73,13 @@ def score(self, X=None, y=None):
             for the outputs relevant to this model type.
 
         Args:
-            X (default None):
-                Ignored.
+            X (bigframes.dataframe.DataFrame | bigframes.series.Series | None):
+                DataFrame of shape (n_samples, n_features). Test samples.
+
+            y (bigframes.dataframe.DataFrame | bigframes.series.Series | None):
+                DataFrame of shape (n_samples,) or (n_samples, n_outputs). True
+                labels for `X`.
 
-            y (default None):
-                Ignored.
         Returns:
             bigframes.dataframe.DataFrame: DataFrame that represents model metrics.
         """