diff --git a/paddle/phi/api/yaml/legacy_api.yaml b/paddle/phi/api/yaml/legacy_api.yaml
index 464ea1f00ba96eb9b727cc0c9f086860d45b2197..6408b908130250bd770e824478ea12f01a909f79 100644
--- a/paddle/phi/api/yaml/legacy_api.yaml
+++ b/paddle/phi/api/yaml/legacy_api.yaml
@@ -922,6 +922,18 @@
   kernel :
     func : greater_than
 
+# grid sample
+- api : grid_sample
+  args : (Tensor x, Tensor grid, str mode, str padding_mode, bool align_corners)
+  output : Tensor(out)
+  infer_meta :
+    func : GridSampleBaseInferMeta
+    param : [x, grid]
+  kernel:
+    func : grid_sample
+    data_type : x
+  backward : grid_sample_grad
+
 - api : group_norm
   args : (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout)
   output : Tensor(y), Tensor(mean), Tensor(variance)
diff --git a/paddle/phi/api/yaml/legacy_backward.yaml b/paddle/phi/api/yaml/legacy_backward.yaml
index 423958399a529970376ee8758b5a37fe4aaead1e..ba8c78edd97ca6c8f06ed065c7f90978f1a290f2 100644
--- a/paddle/phi/api/yaml/legacy_backward.yaml
+++ b/paddle/phi/api/yaml/legacy_backward.yaml
@@ -848,6 +848,18 @@
     data_type : out_grad
   optional: out, dst_count
 
+# grid sample
+- backward_api : grid_sample_grad
+  forward : grid_sample (Tensor x, Tensor grid, str mode, str padding_mode, bool align_corners) -> Tensor(out)
+  args : (Tensor x, Tensor grid, Tensor out_grad, str mode, str padding_mode, bool align_corners)
+  output : Tensor(x_grad), Tensor(grid_grad)
+  infer_meta : 
+    func : GeneralBinaryGradInferMeta
+    param : [x, grid]
+  kernel : 
+    func : grid_sample_grad
+    data_type : x
+
 - backward_api : group_norm_grad
   forward : group_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout) -> Tensor(y), Tensor(mean), Tensor(variance)
   args : (Tensor x, Tensor scale, Tensor bias, Tensor y, Tensor mean, Tensor variance, Tensor y_grad, float epsilon, int groups, str data_layout)
diff --git a/python/paddle/fluid/tests/unittests/test_grid_sampler_op.py b/python/paddle/fluid/tests/unittests/test_grid_sampler_op.py
index f32387d07a9fd142a42910cd59560dd3e0f37ac4..4b813330b405f9e092075d85d85317fbaf12a2e7 100644
--- a/python/paddle/fluid/tests/unittests/test_grid_sampler_op.py
+++ b/python/paddle/fluid/tests/unittests/test_grid_sampler_op.py
@@ -144,6 +144,7 @@ class TestGridSamplerOp(OpTest):
         self.use_cudnn = False
         self.numeric_grad_delta = 0.0001
         self.op_type = 'grid_sampler'
+        self.python_api = paddle.nn.functional.grid_sample
         self.align_corners = True
         self.padding_mode = "zeros"
         self.mode = "bilinear"
@@ -171,13 +172,14 @@ class TestGridSamplerOp(OpTest):
         }
 
     def test_check_output(self):
-        self.check_output()
+        self.check_output(check_eager=True)
 
     def test_check_grad_normal(self):
         self.check_grad(['X', 'Grid'],
                         'Output',
                         max_relative_error=0.01,
-                        numeric_grad_delta=self.numeric_grad_delta)
+                        numeric_grad_delta=self.numeric_grad_delta,
+                        check_eager=True)
 
     def initTestCase(self):
         self.x_shape = (2, 3, 8, 8)
diff --git a/python/paddle/nn/functional/vision.py b/python/paddle/nn/functional/vision.py
index 521a44f758b862a039c195328f95ade2305b5df1..b1a962845c4c54078056a56cbf4064187cb162ec 100644
--- a/python/paddle/nn/functional/vision.py
+++ b/python/paddle/nn/functional/vision.py
@@ -21,6 +21,7 @@ import numpy as np
 from paddle import _C_ops
 from ...device import is_compiled_with_rocm
 from paddle import in_dynamic_mode
+from paddle.fluid.framework import in_dygraph_mode
 from paddle.framework import _non_static_mode
 
 __all__ = []
@@ -272,7 +273,10 @@ def grid_sample(x,
         x.stop_gradient = False
         grid.stop_gradient = False
 
-    if in_dynamic_mode():
+    if in_dygraph_mode():
+        return _C_ops.final_state_grid_sample(x, grid, mode, padding_mode,
+                                              align_corners)
+    elif in_dynamic_mode():
         attrs = ('mode', mode, 'padding_mode', padding_mode, 'align_corners',
                  align_corners, 'use_cudnn', use_cudnn)
         out = getattr(_C_ops, 'grid_sampler')(x, grid, *attrs)