Data analysis technology can deeply analyze and fully mine the physical measurement data of college students to extract valuable information, thus providing data reference for teachers to improve teaching effectiveness. The article first researches the algorithmic process of the Apriori association rule algorithm, combines the transaction compression and hash technology and the Apriori algorithm to further optimize and improve Apriori, and finally applies the mechanism based on the improved data analysis technology in college sports teaching. This paper uses the improved Apriori algorithm to analyze physical test data of students in a college. In the 2021-2022 association rule, data mining results found that the “total score grade” passing students accounted for 74% of the students tested in that year, which can be obtained, the majority of the student’s physical test scores for the passing grade. After a period of a teaching experiment, the p-value of the four dimensions of students’ learning interest is 0.015, 0.048, 0.014, and 0.000, respectively, which is significantly different, thus indicating that the experimental group of students’ learning interest is significantly better than the control group.