The article Aims to compare the application effect of different machine learning algorithms in the recognition of badminton stroke types, so as to better meet the needs of quantitative analysis of badminton skills and tactics, and provide a more reasonable basis for the layout of game strategies.