In power communication networks, load balancing can reduce the overload of bottleneck nodes, improve the utilization of network resources and system reliability.Aiming at the low efficiency of routing algorithm in power communication network, this paper combines variable equivalent multipath algorithm VECMP and improved genetic algorithm IGA for routing optimization in power network based on traditional ECMP algorithm and genetic algorithm.By optimizing three genetic operators, the efficiency of VECMP algorithm is improved, and the superiority of the algorithm in route optimization is verified by comparison with existing optimization algorithms. The results show that the optimization algorithm can quickly obtain the optimal weight allocation scheme of the network and reduce the maximum utilization of the link. This study provides reference and reference for the development of routing optimization methods for power communication networks in China.