This paper proposed a novel hybrid hierarchical strategy for electric vehicle (EV) charging. By synthesizing the strength of the two mainstream control strategies-centralized and decentralized control strategy, it divides the charging process into two levels with different optimal objectives. Considering the stochastic of the charging behavior, it deploys MATSim to simulate the driving patterns and aggregate the EV with similar driving patterns into a so-called Virtual Battery Aggregation (VBA) Model. To verify the effectiveness of the proposed method, an 18-bus system is simulated and 4 different control methods are compared and analysed regarding to peak hour node voltage, line load, and daily load curve.