The propagation delay of cloud computing framework cannot meet the requirements of low delay and reliability of power system. Based on the framework of mobile edge computing, an improved ant colony algorithm for power task scheduling is proposed. On the basis of the shortest path selected by ant colony algorithm, load balancing ability is added to reduce the overall energy consumption and prevent overloading of a mobile edge cloud. The superiority of this scheduling method is verified by simulation analysis. The results show that with the increase of user equipment, edge cloud, bandwidth and total computing resources, the scheduling method proposed in this paper outperforms greedy algorithm and distance priority algorithm in terms of average energy consumption and number of task unloads, which has practical value.