Abstract: The random charging behavior of large-scale electric vehicles not only deteriorates the time distribution characteristics of the load but also has negative impacts on the safe operation of the power grid under the high simultaneity factor of "peak-on-peak" and "power shock". This may lead to suboptimal siting of electric vehicle charging stations. In view of this, this study proposes a multi-objective planning method for electric vehicle charging stations that explicitly considers vehicle-grid interaction. First, by using the utility theory and considering the balance between grid supply and user demand, a vehicle-to-grid interactive response model is constructed. Then, the demand-responde strategy is employed using the Monte Carlo simulation algorithm to enhance the accuracy of electric vehicle charging load forecasting. Consequently, a predictive model that incorporates vehicle-to-grid interaction dynamics is developed. Finally, considering that electric vehicles' participation in the vehicle-to-grid interaction possesses load regulation, peak shifting and valley filling capabilities, a two-tiered multi-objective planning model is established. The upper-layer model aims to minimize comprehensive investment cost of electric vehicle charging infrastructure, while the lower-layer model aims to maximize the charging and discharge benefits of electric vehicle users and the operators. The two-tiered model is solved iteratively using the non-dominated sorting genetic algorithm with elite strategy. Case study results validate the methodological effectiveness and operational feasibility of the developed framework.