With the continuous growth of power industry data, more and more energy is consumed by power dispatching in cloud environment, which aggravates the energy crisis and environmental pollution. Based on the power cloud platform architecture, this paper proposes a power task energy-saving scheduling algorithm based on cloud computing. The virtual machine is assigned to the node with the optimal performance power ratio, and the resource integration is realized by migrating the virtual machine. Through simulation, the scheduling algorithm is compared with the energy-saving virtual machine scheduling node algorithm and the improved best fit decline algorithm. The simulation results show that this method can save more than 10% energy without significantly reducing the efficiency. This study provides a reference for the development of optimal power dispatching method in cloud environment. This work is supported by the Science and Technology Project of State Grid Corporation of China (No. 5212S0170002).