diff --git a/python/paddle/v2/fluid/clip.py b/python/paddle/v2/fluid/clip.py
index b1fd1c2b65f10010fa959dbb47b3fbab114db2f2..eb75018d7798fc326bceba9c00d4fb14aa521d48 100644
--- a/python/paddle/v2/fluid/clip.py
+++ b/python/paddle/v2/fluid/clip.py
@@ -77,6 +77,18 @@ class GradientClipByValue(BaseGradientClipAttr):
         return param, new_grad
 
 
+class GradientClipByNorm(BaseGradientClipAttr):
+    def __init__(self, clip_norm):
+        self.clip_norm = clip_norm
+
+    def process_context(self, context, p_g):
+        pass
+
+    def create_operators(self, param, grad):
+        new_grad = layers.clip_by_norm(x=grad, max_norm=self.clip_norm)
+        return param, new_grad
+
+
 def append_gradient_clip_ops(param_grad):
     context = dict()
     create_op_callbacks = []
diff --git a/python/paddle/v2/fluid/layers/ops.py b/python/paddle/v2/fluid/layers/ops.py
index d3a5b70785947148d6e208b4d8dafec8bb52ff85..884e84011d960cead150dffc51123bd98a7587b5 100644
--- a/python/paddle/v2/fluid/layers/ops.py
+++ b/python/paddle/v2/fluid/layers/ops.py
@@ -16,6 +16,7 @@ __all__ = [
     'elementwise_sub',
     'elementwise_mul',
     'clip',
+    'clip_by_norm',
     'sequence_softmax',
 ] + __activations__