Abstract: The power output of photovoltaic arrays exhibits multi-peak characteristics due to the influence of dust and uneven illumination under actual working conditions. The traditional Maximum Power Point Tracking（MPPT）algorithm cannot achieve global optimization and cannot accurately track to the maximum power point. The genetic algorithm can effectively solve the problem of multi-peak optimization,but the algorithm has problems such as premature and low accuracy in the tracking process. For this reason,a Multiple Population Genetic Algorithm（MPGA）combined with disturbance method is proposed in this paper. In the photovoltaic cell topology model,the optimized dual diode is used instead of the single diode model,and the modeling and simulation are performed under Matlab/Simulink. The results show that the algorithm can accurately,quickly and efficiently find the maximum power point of the photovoltaic array under partial shadow conditions.