diff --git a/backends/xnnpack/partition/config/generic_node_configs.py b/backends/xnnpack/partition/config/generic_node_configs.py
index b95d7c5b89..f08b8ccb3c 100644
--- a/backends/xnnpack/partition/config/generic_node_configs.py
+++ b/backends/xnnpack/partition/config/generic_node_configs.py
@@ -303,6 +303,21 @@ def get_original_aten(self) -> Optional[torch._ops.OpOverload]:
 class UpsampleBilinear2dConfig(GenericNodePartitionerConfig):
     target_name = "upsample_bilinear2d.vec"
 
+    def check_constraints(self, node: torch.fx.Node, ep: ExportedProgram) -> bool:
+        """
+        XNNPACK's static_resize_bilinear does not support dynamic output sizes
+        """
+        if not self.check_common_constraints(node, ep):
+            return False
+
+        is_output_dynamic = "val" in node.meta and any(
+            isinstance(d, torch.SymInt) for d in node.meta["val"].shape
+        )
+        if is_output_dynamic:
+            why(node, reason="dynamic output sizes are not supported")
+            return False
+        return True
+
     def supported_precision_types(self) -> List[ConfigPrecisionType]:
         return [ConfigPrecisionType.FP32]
 
diff --git a/backends/xnnpack/test/ops/bilinear2d.py b/backends/xnnpack/test/ops/bilinear2d.py
index bf89e2196f..6a19476365 100644
--- a/backends/xnnpack/test/ops/bilinear2d.py
+++ b/backends/xnnpack/test/ops/bilinear2d.py
@@ -8,6 +8,7 @@
 
 import torch
 from executorch.backends.xnnpack.test.tester import Tester
+from executorch.backends.xnnpack.test.tester.tester import Export
 
 
 class TestUpsampleBilinear2d(unittest.TestCase):
@@ -118,3 +119,23 @@ def test_fp32_static_resize_bilinear2d_antialiased(self):
             )
             .check_not(["torch.ops.higher_order.executorch_call_delegate"])
         )
+
+    def test_fp32_bilinear2d_dynamic_bilinear2d_not_partitioned(self):
+        """
+        Verify that upsample_bilinear2d ops with dynamic output sizes are not partitioned.
+        """
+        example_inputs = (torch.randn(2, 3, 4, 5),)
+        dynamic_shapes = {
+            "x": {
+                2: torch.export.Dim("h", min=1, max=10),
+                3: torch.export.Dim("w", min=1, max=12),
+            }
+        }
+        (
+            Tester(self.StaticResizeBilinear2dModule(), example_inputs)
+            .export(Export(dynamic_shapes))
+            .to_edge_transform_and_lower()
+            # NOTE The decomposition is partially delegated. This will need to be replaced
+            # with the aten upsample op once decomp is removed.
+            .check("executorch_exir_dialects_edge__ops_aten_index_Tensor")
+        )