This paper focuses on the multi-dimensional characteristics of college music teachers’ teaching ability, and constructs a four-dimensional evaluation index system containing professional ethics, practical ability, teaching and research ability, and expansion ability. In order to solve the multi-objective collaborative optimization problem, the constrained particle swarm algorithm (TBC-PSO) based on two-stage adaptive angular region division is proposed, which divides the whole optimization process into two stages of adaptive switching, and adopts different optimization strategies respectively. Through the balance of inter-group homogeneity and intra-group heterogeneity, the precise design of teaching ability improvement strategy is realized. The weights of the evaluation index system are determined, and a 16-week blended teaching experiment is carried out with a sample of 15 music teachers in a provincial university. The average value of teachers’ teaching ability scores increased from 75.04 to 80.53 after the experiment, and all teachers’ teaching ability scores exceeded 78, and 9 out of 15 teachers’ comprehensive scores increased by more than 5 points, which verified the effectiveness of this paper’s optimization scheme.
