In recent years, the penetration rate of electric vehicles in the distribution network has gradually increased, which leads to the increasingly obvious difference of power flow distribution in the distribution network. The congestion management of the distribution network is a necessary link of the smart grid. In this paper, considering the elasticity of flexible loads, a decentralized scheduling strategy for electric vehicles based on virtual aggregation and AC optimal power flow （ACOPF） is proposed. Firstly, starting from the synergistic relationship of multiple market entities, a framework for flexible load participation of electric vehicles in distribution network scheduling is designed. Secondly, a virtual aggregation model of electric vehicles is established based on the charging sequence relationship in the electric vehicle station. Then considering the load elasticity, the two-layer optimization model of market economy safety dispatching and in-station electric vehicle dispatching is established based on ACOPF. Finally, the simulation results show that the proposed strategy can improve the power flow distribution of the distribution network and improve the solution efficiency. What’s more, the node marginal electricity price can be used as a fair price signal to guide electric vehicles charging in an orderly manner, enabling decentralized dispatching of electric vehicles and reducing costs.