Abstract: The aim of this research is to discuss how to optimize Model Reference Adaptive Control (MRAC) for a Permanent Magnet Synchronous Motor (PMSM). With the help of MRAC that can overcome the influence of change of motor parameters and load, PMSM is characterized by good response and robustness. Generally speaking, PMSM controller has a long-proven infrastructure without complex computation, of which the Small-the-Better (STB) output features of PMSM include: Overshoot, Rise Time and Settling Time. In previous designs of controllers, only individual quality characteristics were considered without overall output planning of multi-quality characteristics. By combining quality characteristics into a single quality characteristic index through Desirability Function, this paper strives for optimization of parameters using Genetic Algorithm (GA), thereby improving the robustness and output performance of PMSM speed control. By using directly the stator current and speed signals in the control structure, a system integrating hardware and software was developed through PC-based motor controller. The results from simulation and experiment show that, under different rotational speeds with or without load, the adaptive control policy could offer better Multi-response performance.