With the increase in scale of power consumption in low-voltage distribution networks, three-phase imbalance problems are becoming more increasingly serious. This leads to increased losses and the safety problems of equipment. Use of a commutation switch is the fundamental way to solve this problem, but the current scheme only considers the amplitude imbalance. Thus its calculation results are not sufficiently comprehensive nor accurate. This paper proposes a type of intelligent commutation switch strategy, in which the amplitude and phase of current imbalance are considered simultaneously, and the current imbalance degree and the number of commutation switch actions are taken as the multi-objective function, and the mutation probability based on a vector gene genetic algorithm is improved. This determines the optimal commutation switch combination more quickly. Finally, the effectiveness of the proposed control strategy is verified by Matlab/Simulink experiment. The experimental results show that the intelligent switching strategy can effectively improve the three-phase imbalance in a distribution area. This work is supported by the National Natural Science Foundation of China (No. 51877141).