diff --git a/python/paddle/fluid/tests/unittests/test_clip_op.py b/python/paddle/fluid/tests/unittests/test_clip_op.py
index 74c5f693a37f1f0a480a14465e87be97cabb8f9f..f4423ccd0294cc8cdf88a20b75c9f6f32dddd3db 100644
--- a/python/paddle/fluid/tests/unittests/test_clip_op.py
+++ b/python/paddle/fluid/tests/unittests/test_clip_op.py
@@ -20,11 +20,13 @@ import paddle
 import paddle.fluid as fluid
 from paddle.fluid import Program, program_guard
 from op_test import OpTest
+from paddle.fluid.framework import _test_eager_guard
 
 
 class TestClipOp(OpTest):
     def setUp(self):
         self.max_relative_error = 0.006
+        self.python_api = paddle.clip
 
         self.inputs = {}
         self.initTestCase()
@@ -51,12 +53,12 @@ class TestClipOp(OpTest):
 
     def test_check_output(self):
         paddle.enable_static()
-        self.check_output()
+        self.check_output(check_eager=True)
         paddle.disable_static()
 
     def test_check_grad_normal(self):
         paddle.enable_static()
-        self.check_grad(['X'], 'Out')
+        self.check_grad(['X'], 'Out', check_eager=True)
         paddle.disable_static()
 
     def initTestCase(self):
@@ -228,6 +230,10 @@ class TestClipAPI(unittest.TestCase):
         self.assertTrue(
             np.allclose(out_5.numpy(), (data * 10).astype(np.int64).clip(2, 8)))
 
+    def test_eager(self):
+        with _test_eager_guard():
+            self.test_clip_dygraph()
+
     def test_errors(self):
         paddle.enable_static()
         x1 = fluid.data(name='x1', shape=[1], dtype="int16")
diff --git a/python/paddle/tensor/math.py b/python/paddle/tensor/math.py
index d2ed985fb86516e5f78c953c769fc53f32c47de9..e4faa573ffb265f00cf2c1961ecd56bdfdb9ce35 100755
--- a/python/paddle/tensor/math.py
+++ b/python/paddle/tensor/math.py
@@ -2290,7 +2290,16 @@ def clip(x, min=None, max=None, name=None):
         min_ = float(np.finfo(np.float32).min)
         max_ = float(np.finfo(np.float32).max)
 
-    if paddle.in_dynamic_mode():
+    if in_dygraph_mode():
+        if isinstance(min, Variable):
+            min = min.numpy().item(0)
+        if isinstance(max, Variable):
+            max = max.numpy().item(0)
+        min = min_ if min is None else min
+        max = max_ if max is None else max
+        return _C_ops.final_state_clip(x, min, max)
+
+    if _in_legacy_dygraph():
         if isinstance(min, Variable):
             min = min.numpy().item(0)
         if isinstance(max, Variable):
@@ -2350,7 +2359,12 @@ def clip_(x, min=None, max=None, name=None):
         max = max.numpy().item(0)
     min = fmin if min is None else min
     max = fmax if max is None else max
-    return _C_ops.clip_(x, "min", min, "max", max)
+
+    if in_dygraph_mode():
+        return _C_ops.final_state_clip_(x, min, max)
+
+    if _in_legacy_dygraph():
+        return _C_ops.clip_(x, "min", min, "max", max)
 
 
 
diff --git a/python/paddle/utils/code_gen/api.yaml b/python/paddle/utils/code_gen/api.yaml
index 4c17644792fbdeee9e19559d6f50b801be6303ef..08cf04f692806d720325bae56aaa127b49c071da 100644
--- a/python/paddle/utils/code_gen/api.yaml
+++ b/python/paddle/utils/code_gen/api.yaml
@@ -312,6 +312,17 @@
     func : cholesky_solve
   backward : cholesky_solve_grad
 
+- api : clip
+  args : (Tensor x, Scalar(float) min, Scalar(float) max)
+  output : Tensor(out)
+  inplace : (x -> out)
+  infer_meta :
+    func : UnchangedInferMeta
+    param : [x]
+  kernel :
+    func : clip
+  backward : clip_grad
+
 - api : concat
   args : (Tensor[] x, Scalar(int64_t) axis)
   output : Tensor
diff --git a/python/paddle/utils/code_gen/backward.yaml b/python/paddle/utils/code_gen/backward.yaml
index da60dae43169542909a4d108d95630eeaf3aa635..570e64dcd5e12eeedf199398becbf0503ba0d89b 100644
--- a/python/paddle/utils/code_gen/backward.yaml
+++ b/python/paddle/utils/code_gen/backward.yaml
@@ -179,6 +179,16 @@
   kernel :
     func : cholesky_solve_grad
 
+- backward_api : clip_grad
+  forward : clip (Tensor x, Scalar min, Scalar max) -> Tensor(out)
+  args : (Tensor x, Tensor out_grad, Scalar min = 0., Scalar max = 0.)
+  output : Tensor(x_grad)
+  infer_meta :
+    func : UnchangedInferMeta
+    param : [x]
+  kernel :
+    func : clip_grad
+
 - backward_api : concat_grad
   forward : concat (Tensor[] x, Scalar axis) -> Tensor(out)
   args : (Tensor[] x, Tensor out_grad, Scalar axis = 0)