The P-U characteristics of the PV array with bypass diodes have multi local maximum power points under partially shadowed conditions, but Conventional Maximum Power Point Tracking (CMPPT) methods may fail to identify global maximum power points. The number of the extreme points in the output characteristics of PV array is studied, and the artificial bee colony algorithm based on differential evolution is applied to the MPPT. First, the initial positions of the bees are pre-defined to avoid missing the extreme points. Combining the mutation strategy in the differential evolution algorithm with the artificial bee colony algorithm to realize the global maximum power point tracking control under time-varying conditions, and efficient iteration stop strategy is proposed which could reduce the power oscillation when trending to steady state. The S-Function model is built in Matlab to verify the effectiveness of the algorithm.