In this paper, multiobjective design optimization (MDO) of three-phase permanent magnet synchronous machines is presented which has been formulated from the analytical sub-domain model and particle swarm optimization. Five machine parameters i.e. magnet arc, slot opening, magnet thickness, airgap length and stator inner radius, are identified in the analytical sub-domain model and then optimized using the particle swarm optimization (PSO) algorithm to provide the optimal motor design with certain objective functions. The objective functions are the low cogging torque, low total harmonics distortion, high output torque, and high efficiency. The results from this MDO show that the magnet arc is optimized from full-pitch to 0.8 of full-pitch which yields 20% cost-saving in magnet volume, the back-emf total harmonics distortion of 4.7%, the cogging torque of 0.45Nm and 93.1% efficiency when the stator inner radius, Rsi is set to 32mm for 15slot/10pole PMSM. Whereas, the 6slot/4pole PMSM exhibits the back-emf total harmonics distortion of 9.3%, the cogging torque of 0.17Nm and 95.7% efficiency when the stator inner radius, Rsi is set to 20.9mm. MDO which combines the analytical sub-domain model with PSO algorithm can lead to faster PMSM design process and the results are comparable with that of finite element method.