diff --git a/python/paddle/fluid/tests/book/notest_understand_sentiment.py b/python/paddle/fluid/tests/book/notest_understand_sentiment.py
index 241778e303036d068dc0a40e4574a02eb97ad134..792ed7368d646cd9dff9255eb402b6a9b84f69a6 100644
--- a/python/paddle/fluid/tests/book/notest_understand_sentiment.py
+++ b/python/paddle/fluid/tests/book/notest_understand_sentiment.py
@@ -170,7 +170,7 @@ def train(word_dict,
         assert save_dirname is None
 
     adagrad = fluid.optimizer.Adagrad(learning_rate=0.002)
-    optimize_ops, params_grads = adagrad.minimize(cost)
+    adagrad.minimize(cost)
 
     train_data = paddle.batch(
         paddle.reader.shuffle(
diff --git a/python/paddle/fluid/tests/book/test_fit_a_line.py b/python/paddle/fluid/tests/book/test_fit_a_line.py
index ecb34699af0dc14782601702ab8afedbca7e1bfd..b1a6b524d33cae97c8982ffb8f780b1b07761a09 100644
--- a/python/paddle/fluid/tests/book/test_fit_a_line.py
+++ b/python/paddle/fluid/tests/book/test_fit_a_line.py
@@ -33,7 +33,7 @@ def train(use_cuda, save_dirname, is_local):
     avg_cost = fluid.layers.mean(cost)
 
     sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
-    optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)
+    sgd_optimizer.minimize(avg_cost)
 
     BATCH_SIZE = 20
 
diff --git a/python/paddle/fluid/tests/book/test_image_classification.py b/python/paddle/fluid/tests/book/test_image_classification.py
index dbcdb5766e7d20efdb12da0ea4c6f005d903849b..0f3a4c9242a81a3c1fb90268245715a8e59a207a 100644
--- a/python/paddle/fluid/tests/book/test_image_classification.py
+++ b/python/paddle/fluid/tests/book/test_image_classification.py
@@ -125,7 +125,7 @@ def train(net_type, use_cuda, save_dirname, is_local):
     test_program = fluid.default_main_program().clone(for_test=True)
 
     optimizer = fluid.optimizer.Adam(learning_rate=0.001)
-    optimize_ops, params_grads = optimizer.minimize(avg_cost)
+    optimizer.minimize(avg_cost)
 
     BATCH_SIZE = 128
     PASS_NUM = 1
diff --git a/python/paddle/fluid/tests/book/test_label_semantic_roles.py b/python/paddle/fluid/tests/book/test_label_semantic_roles.py
index 0faba33032d5dfc0b751a5191e7b2ae0c1f172bf..09793760e5504c04ad4b0bfac5c5d7b7047cf85d 100644
--- a/python/paddle/fluid/tests/book/test_label_semantic_roles.py
+++ b/python/paddle/fluid/tests/book/test_label_semantic_roles.py
@@ -175,7 +175,7 @@ def train(use_cuda, save_dirname=None, is_local=True):
             decay_steps=100000,
             decay_rate=0.5,
             staircase=True))
-    optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)
+    sgd_optimizer.minimize(avg_cost)
 
     # TODO(qiao)
     # add dependency track and move this config before optimizer
diff --git a/python/paddle/fluid/tests/book/test_machine_translation.py b/python/paddle/fluid/tests/book/test_machine_translation.py
index 46c6b9c29a265741a99655d5ac29244798f6fec2..e8a75f473f62df528b7f39bf5f9085076e005c25 100644
--- a/python/paddle/fluid/tests/book/test_machine_translation.py
+++ b/python/paddle/fluid/tests/book/test_machine_translation.py
@@ -185,7 +185,7 @@ def train_main(use_cuda, is_sparse, is_local=True):
         learning_rate=1e-4,
         regularization=fluid.regularizer.L2DecayRegularizer(
             regularization_coeff=0.1))
-    optimize_ops, params_grads = optimizer.minimize(avg_cost)
+    optimizer.minimize(avg_cost)
 
     train_data = paddle.batch(
         paddle.reader.shuffle(
diff --git a/python/paddle/fluid/tests/book/test_recognize_digits.py b/python/paddle/fluid/tests/book/test_recognize_digits.py
index c115aa4d7d6b514f9207543730e5e76cb0d2040c..578b1162fbd7e3a1b1c0cc934406818f2e07e019 100644
--- a/python/paddle/fluid/tests/book/test_recognize_digits.py
+++ b/python/paddle/fluid/tests/book/test_recognize_digits.py
@@ -95,7 +95,7 @@ def train(nn_type,
     test_program = fluid.default_main_program().clone(for_test=True)
 
     optimizer = fluid.optimizer.Adam(learning_rate=0.001)
-    optimize_ops, params_grads = optimizer.minimize(avg_loss)
+    optimizer.minimize(avg_loss)
 
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
diff --git a/python/paddle/fluid/tests/book/test_recommender_system.py b/python/paddle/fluid/tests/book/test_recommender_system.py
index d022dedbff805d597b68b5a47f7931f2dd946615..7be924f762ddeb045dda890dbfdcd96a65449553 100644
--- a/python/paddle/fluid/tests/book/test_recommender_system.py
+++ b/python/paddle/fluid/tests/book/test_recommender_system.py
@@ -160,7 +160,7 @@ def train(use_cuda, save_dirname, is_local=True):
     test_program = fluid.default_main_program().clone(for_test=True)
 
     sgd_optimizer = SGDOptimizer(learning_rate=0.2)
-    optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)
+    sgd_optimizer.minimize(avg_cost)
 
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
diff --git a/python/paddle/fluid/tests/book/test_word2vec.py b/python/paddle/fluid/tests/book/test_word2vec.py
index 6dec0f6857e86b4b9c1c67af934aa9bfdb1c3df7..30e1a5040cc92b02bbbf90dac97001812ec90134 100644
--- a/python/paddle/fluid/tests/book/test_word2vec.py
+++ b/python/paddle/fluid/tests/book/test_word2vec.py
@@ -101,7 +101,7 @@ def train(use_cuda, is_sparse, is_parallel, save_dirname, is_local=True):
         avg_cost = fluid.layers.mean(pd())
 
     sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
-    optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)
+    sgd_optimizer.minimize(avg_cost)
 
     train_reader = paddle.batch(
         paddle.dataset.imikolov.train(word_dict, N), BATCH_SIZE)