Abstract: Traditional methods for the extraction and identification of the data of dynamic isolated gestures, which constitute a complex and high cost, which is easy to be disturbed by the external environment, and the recognition accuracy is low. The study of dynamic hand gesture recognition method based on Kinect, obtaining the depth information of dynamic gesture using Kinect sensor, the human hand positioning real-time and accurate tracking, feature and image denoising processing, the extraction of hand movement; finally introduce the hidden Markov model (HMM) for effective training on dynamic gesture and moving track sample set, to achieve dynamic hand gesture recognition. Experiments show that the proposed method has high recognition rate and high robustness in the environment of noise interference and light loss.