This study is based on deep learning algorithm wushu movement recognition and optimization methods, in order to improve the accuracy and efficiency of wushu movement recognition. Martial arts movement recognition is a challenging task because of the wide variety of martial arts movements, the subtle differences between movements, and the continuity and speed changes of movements also increase the difficulty of recognition. This paper proposes a martial arts movement recognition and optimization method based on a deep learning algorithm to improve the accuracy and efficiency of martial arts movement recognition. This paper introduces the background and significance of wushu movement recognition and the application of deep learning algorithms in wushu movement recognition. Then, this paper describes the optimization methods in detail, including data set construction, feature extraction, model training and optimization. In this paper, the experimental results are analyzed in detail, and the experimental results are compared and discussed. This paper summarizes the research on wushu movement recognition and optimization methods and puts forward a new optimization method to improve the performance of movement recognition.