Against the background of physical education teaching reform in universities across the country, the traditional teaching model has revealed its limitations. Teacher-centered instruction, insufficient personalized guidance for students, passive learning, and low efficiency in practical application have become urgent problems to be solved in current basketball course teaching. Based on the traditional “explanation-demonstration-practice” structure, it is found in the 90-minute class that: the frequency of teacher-student interaction is 0.8 times per minute, the actual effective practice time of students is about 20%, the average concentration duration of students is generally no more than 17 minutes, less than 45% of the skills learned by students in traditional teaching can be applied in actual combat, and there is an obvious difference in students’ sports ability in teaching classes (Gini coefficient reaches 0.49). Traditional teaching usually sets teaching objectives for students with medium sports ability. In view of the above phenomena, we attempt to construct a new hierarchical teaching method based on the “micro-competition” teaching model, relying on current big data platforms and advanced multimedia technologies. We organize small groups (such as 2v2, 3v3, 4v4, 5v5) to carry out high-frequency and short-duration confrontations, and provide hierarchical guidance on students’ techniques and tactics. To realize precise grouping and personalized guidance, the TTPQ four-dimensional evaluation system is introduced. Emphasizing students’ dominance in competitions, we aim to improve students’ abilities in technique and tactic application, timely decision-making and team cooperation in actual combat through tactical design, role rotation and real-time review and improvement, so as to provide a feasible path for changing the current situation of “unified teaching, neglecting actual combat” in university physical education.