The ever-evolving landscape of educational technology presents new challenges and opportunities for university sports practical teaching. Traditional teaching methodologies, while emphasizing standardization in form, structure, and content, often struggle to cater to individual student needs and enhance instructional efficacy. The Classification and Regression Trees (CART) algorithm, renowned for its robust classification and prediction capabilities, has found widespread application across various domains. This paper delves into the potential of applying the CART algorithm in university sports practical teaching, aiming to provide novel insights and approaches for enhancing the quality of sports education. Introduced by Breiman et al. in 1984.The resulting decision tree is capable of handling nonlinear relationships, automatically selecting features, and detecting interactions. In the context of classification problems, CART excels, offering a powerful tool for educational data analysis. A comprehensive analysis of the research results shows that there are significant deviations in the process of higher education governance. The method proposed by me can optimize the existing teaching governance content and achieve comprehensive governance of physical education teaching.