In order to reduce the permanent magnet cost and improve the performance of external rotor permanent magnet synchronous motor (ERPMSM) for belt conveyor, an optimization method based on response surface methodology (RSM) and improved multi-objective particle swarm optimization (IMOPSO) algorithm is proposed. Based on the establishment of the motor’s basic structure, the permanent magnet size, the air gap length and the slot width are taken as the optimization parameters, and the permanent magnet cost, the output torque, and the torque ripple are taken as the optimization objectives. The significant parameters are screened by sensitivity analysis. The sample space is established based on RSM and finite element simulation, the function relationship between optimization objectives and parameters is fitted. The scheme is optimized by IMOPSO. Finally, the results of comparison of the schemes before and after optimization show that the multi-objective optimization algorithm is accurate and reliable, it has better convergence and diversity, which can optimize the performance of the motor while reducing the cost of permanent magnet simultaneously.