Abstract: In order to make full use of the river resources and enhance the system stability, a research method for operation optimization control of the diversion power generation system of hydropower station considering tailwater excavation is proposed to solve the problems of mixed operating conditions, frequent oscillations and weak stability. This method is based on PI governor of hydropower station, and simulates the characteristics of firefly to attract the same species and fly to the brighter groups. The weighted sum of absolute values of maximum and minimum speed of the simulation system is taken as the optimization objective function, and the proportional and integral coefficient of the governor are optimized by firefly algorithm. By applying load increase and reduction conditions, the governor control performance of hydropower station diversion power generation system is compared with different tailwater excavation depth. Experimental data show that the proposed method can achieve better stability control performance than the method of unoptimized and Stein optimized governor, and the damping oscillation characteristics of hydropower station diversion power generation system are significantly improved. The governor optimized by firefly algorithm has good robust performance under different load increase and reduction conditions, and the regulation quality of unit speed has significantly improved.