A fast and reliable protection can be provided by the Quench Protection Unit(QPU) for preventing the damage of superconducting magnets. According to the design of QPU in International Thermonuclear Experimental Reactor(ITER), when quench occurs, the main circuit current is forced to reduce to zero by commutation circuit, which plays the crucial rule in the reliability of QPU. In this paper, the effect of component parameters in commutation circuit on the reliability of QPU is analyzed. In addition, a parameter optimization method based on the Genetic Algorithm(GA), which aims to obtain the optimal reliability of QPU is presented. The correctness of the optimization mothed is verified by simulation, which proves the optimized component parameters obtained by this method have significant effect on enhancing the reliability of the unit to give superconducting magnets a better protection.