This study utilized the Delphi method to construct a four-level indicator system for core competencies in physical education teaching. Through three rounds of consultation involving 20 experts in the field of physical education, the initial 47 third-level indicators were screened and revised, ultimately forming an optimized system comprising 4 first-level indicators, 10 second-level indicators, and 41 third-level indicators. Based on this framework, the Entropy Weight-TOPSIS method was applied to evaluate 92 physical education teacher trainees at a certain university. The Entropy Weight method revealed that the top three third-level indicator weights were C73 ProblemSolving Ability (4.19%), C53 Teaching Implementation Ability (3.98%), and C51 Teaching Demonstration Ability (3.86%). TOPSIS analysis indicated that there were significant differences in student competencies, with the optimal solution (Student30, D⁺=0) and the worst solution (Student46, D⁻=0.0027) differing by a factor of 30 in terms of proximity, and the top 10% of students (S>0.0628) far exceeding the bottom 10% (S<0.0112) in terms of proximity. The article proposes a cultivation path for students' core competencies in physical education based on a mechanism for dynamically adjusting indicator weights, providing a more precise competency cultivation model for physical education integration policies.
