Abstract: In order to improve the flexibility of distribution network planning, this paper proposes a flexible distribution system scalable planning model based on approximate dynamic programming, considering the long-term load growth uncertainty and short-term output fluctuation. Firstly, the current cost and future variable cost are considered, and the Markov decision process based on flexible investment strategy is used to make decisions in each stage of distribution network planning. Secondly, the DSEP model is constructed by minimizing the investment cost, maintenance cost, production cost and energy cost under the condition of load growth, and taking the investment utilization constraint, power flow constraint, radial constraint and upper and lower limit constraint as constraints. Then the approximate dynamic programming method is used to decompose the multi-stage optimization problem into easily solvable sequential subproblems. Finally, the feasibility and superiority of the planning model are verified by the case study of the distribution system with 3 stages and 24 nodes and 5 stages and 54 nodes. The flexible and scalable scheme realizes the reasonable allocation of renewable energy equipment.