Aiming at the problem that the traditional NSGA-Ⅱ has weak regional local search ability in multi-objective optimization of permanent magnet (PM) assisted switched reluctance motors, a regional local search algorithm based on NSGA-Ⅱ (RLS-NSGA-Ⅱ) is proposed. The improved crossover operator and mutation operator are used to enhance the search capability of the algorithm and search for the optimal solution set in the local area. The effectiveness of the algorithm is verified through standard multi-objective problems. Aiming at the problem that the optimization algorithm′s optimization precision of decision variables is too high, effective units for decision variables are reserved, and a minimum unit optimization mutation method is proposed. The multi-objective optimization of motor torque ripple, efficiency and average torque is performed using the proposed algorithm and the traditianal NSGA-Ⅱ. After optimization effect comparison and simulation, it is verified that the optimal solution derived from the proposed algorithm is more advantageous.