With the development of science and technology and artificial intelligence algorithms, sports and sports-related teaching are not improving in a spiral shape. Among them, intelligent video extraction technology and intelligent speech recognition technology are the most widely used. In this paper, based on PCA, LDA and other data dimensionality reduction algorithms, the application of intelligent speech recognition technology in physical education is studied. After studying the feature extraction, the speech and semantic recognition technology improves the recognition of sports actions by 8%, and the results are significantly improved.