Abstract: Flexible interconnection technology is one of the effective means to solve numerous problems faced by distribution networks with high proportion of distributed generations (DGs). A multi-layer optimization based coordinated planning method for medium-voltage and low voltage flexible interconnection of distribution networks is proposed. Firstly, a power flow model for distribution network with medium-voltage and low-voltage flexible interconnection based on power electronics flexible interconnected devices (FID) is built. Then, a three-layer coordinated planning model is constructed. With the upper layer aiming to minimize the annual operating cost of low-voltage FIDs and the annual variance of transformer load rate in substation areas, and the middle layer aiming to minimize the annual operating cost of medium voltage FIDs and the annual electricity purchasing cost from the higher-level power grid, they respectively determine the installation location and capacity of low-voltage and medium-voltage FIDs. The lower layer optimizes the system operation with the goal of minimizing the electricity purchasing cost from the higher-level power grid in each scenario. A hybrid algorithm based on adaptive particle swarm optimization and second-order cone programming is adopted for solution. Finally, an IEEE-33 node distribution network with a high proportion of DGs is used for case study, and the results indicate that through flexible interconnection planning, the annual comprehensive operating cost of the example system is reduced by 19.01% and the annual variance of transformer load rate in substation areas is reduced by 82.59%, showing the effectiveness of the proposed planning model.