Compound MPPT algorithm based on particle swarm optimization was proposed for solving the problem of maximum power point tracking under non-uniform insolation condition. Compound PSO-MPPT control algorithm contained off line optimization and on line optimization. During the process of off line optimization, information exchange mechanism between adjacent particles was used which can increase convergence speed and the search capability for critical areas. Fitness fuzzy judgment method was used to decrease the influence of disturbance and shorten the time-consuming which on line particle swarm optimization was considered. Compound PSO-MPPT control algorithm combined off line optimization and on line optimization together and it made up the shortfall of each method. The compound PSO-MPPT can acquire the faster response speed and higher accuracy of maximum power point than traditional MPPT algorithm. The proposed algorithm is testified by Matlab simulations and can realize the maximum power point tracking under non-uniform insolation condition.