As an important application in computer vision field, human action recognition with depth video sequence comes to a hot research topic in recent years. With the popularity of Kinect devices, which makes the application of action recognition with depth video can be applied to intelligent monitoring more effective. This paper makes full use of the existing Kinect extraction technology and human action recognition technology. Specifically, we extracted human body depth images and skeleton data information from deep data streams obtained by the Kinect SDK in real time. Furthermore, with human action characteristics are extracted, human action recognized can be realized by the way of making comparison for its application in intelligent monitoring. In addition, we verify the proposed method by designing of the intelligent surveillance platform. The experimental results show that the proposed method achieves good performance in the real cheating detection task and has good application value.