diff --git a/paddle/operators/momentum_op.cc b/paddle/operators/momentum_op.cc
index e8ce16f4cfcf83fd13e4d3a5318a4ae0c8c8449c..19954006195c1e9fd34328b52ed2a9eade526235 100644
--- a/paddle/operators/momentum_op.cc
+++ b/paddle/operators/momentum_op.cc
@@ -75,7 +75,7 @@ class MomentumOpMaker : public framework::OpProtoAndCheckerMaker {
     AddOutput("VelocityOut", "(Tensor) Output updated velocity");
 
     AddAttr<float>("mu", "(float) Momentum coefficient");
-    AddAttr<bool>("useNesterov",
+    AddAttr<bool>("use_nesterov",
                   "(bool, default false) "
                   "Use Nesterov Momentum")
         .SetDefault(false);
diff --git a/paddle/operators/momentum_op.h b/paddle/operators/momentum_op.h
index e6d6d1da3df9f7e43a93fcc2e12658a01a491f81..8f7f5eb5c21c0342f57a47b85d28f4454f4566c2 100644
--- a/paddle/operators/momentum_op.h
+++ b/paddle/operators/momentum_op.h
@@ -34,7 +34,7 @@ class MomentumOpKernel : public framework::OpKernel<T> {
     velocity_out->mutable_data<T>(ctx.GetPlace());
 
     float mu = ctx.Attr<float>("mu");
-    bool use_nesterov = ctx.Attr<bool>("useNesterov");
+    bool use_nesterov = ctx.Attr<bool>("use_nesterov");
 
     auto p_out = framework::EigenVector<T>::Flatten(*param_out);
     auto v_out = framework::EigenVector<T>::Flatten(*velocity_out);
diff --git a/python/paddle/v2/framework/optimizer.py b/python/paddle/v2/framework/optimizer.py
index f20865d604f68c8398f299bf8edfd020bfa4e4c5..5b4cdecf2c4285618131657a09fbe437191ea75a 100644
--- a/python/paddle/v2/framework/optimizer.py
+++ b/python/paddle/v2/framework/optimizer.py
@@ -297,7 +297,7 @@ class MomentumOptimizer(Optimizer):
                 "VelocityOut": velocity_acc
             },
             attrs={"mu": self._momentum,
-                   "useNesterov": self._use_nesterov})
+                   "use_nesterov": self._use_nesterov})
 
         return momentum_op
 
diff --git a/python/paddle/v2/framework/tests/test_momentum_op.py b/python/paddle/v2/framework/tests/test_momentum_op.py
index 654d31975aab4578055e7e70ade202bd2c3d93cb..638095f7564c8761151a7794f98f9ca797b0083b 100644
--- a/python/paddle/v2/framework/tests/test_momentum_op.py
+++ b/python/paddle/v2/framework/tests/test_momentum_op.py
@@ -37,7 +37,7 @@ class TestMomentumOp1(OpTest):
 
 
 class TestMomentumOp2(OpTest):
-    '''Test Momentum with defaukt values for attributes
+    '''Test Momentum with default values for attributes
     '''
 
     def setUp(self):
@@ -57,7 +57,7 @@ class TestMomentumOp2(OpTest):
             'LearningRate': learning_rate
         }
 
-        self.attrs = {'mu': mu, 'useNesterov': use_nesterov}
+        self.attrs = {'mu': mu, 'use_nesterov': use_nesterov}
 
         velocity_out = mu * velocity + grad
         if use_nesterov:
diff --git a/python/paddle/v2/framework/tests/test_optimizer.py b/python/paddle/v2/framework/tests/test_optimizer.py
index 9333df8f7f347a080cfb035ccd0c575ded7c423a..a39e7402600c7a94301de030c90ea51264248cf1 100644
--- a/python/paddle/v2/framework/tests/test_optimizer.py
+++ b/python/paddle/v2/framework/tests/test_optimizer.py
@@ -98,7 +98,7 @@ class TestMomentumOptimizer(unittest.TestCase):
         self.assertEqual(len(opts), 1)
         sgd_op = opts[0]
         self.assertEqual(sgd_op.type, "momentum")
-        self.assertFalse(sgd_op.attr('useNesterov'))
+        self.assertFalse(sgd_op.attr('use_nesterov'))
 
         # Check accumulators
         accumulators = momentum_optimizer.get_accumulators()
@@ -143,7 +143,7 @@ class TestMomentumOptimizer(unittest.TestCase):
         self.assertEqual(len(opts), 1)
         sgd_op = opts[0]
         self.assertEqual(sgd_op.type, "momentum")
-        self.assertTrue(sgd_op.attr('useNesterov'))
+        self.assertTrue(sgd_op.attr('use_nesterov'))
 
         # Check accumulators
         accumulators = momentum_optimizer.get_accumulators()