The optimal sizing of microgrid has more uncertain influence with large-scale of electric vehicles(EV) integrated into microgrid and unstable renewable energy sources such as wind and solar energy. The price incentive model and monte carlo method was applied to simulating different management modes of EV load. Based on scenario generation and K-medoids cluster, the sizing optimization would be done by using particle swarm algorithm in typical scenarios with the objective to minimize the comprehensive cost. The simulation result validates that optimal sizing will increase the reliability of microgrid considering randomness of wind and solar energy.And effective energy management mode can decrease the fluctuation of load, reduce the annual costs of EV users and comprehensive costs of microgrid.