In the case of partial shadow, there will be multiple peak points in the P-U curve of the photovoltaic(PV) array, and the conventional maximum power tracking method will fail. This paper proposes a particle swarm optimization algorithm based on adaptive weight. Based on the global optimization characteristics of particle swarm optimization, the multi-peak P-U curve of PV array is optimized under local shadow, and the voltage corresponding to the maximum power point is the best output voltage. Combined with voltage closed-loop regulation and Boost circuit, this paper constructs the photovoltaic system simulation model to simulate the maximum power output, which is compared with traditional perturbation and observation method and simulated in Matlab/Simulink.