In order to solve the problem of nonlinear power flow model modeling and uncertainty optimization efficiency in existing distribution system planning, a chance-constrained optimization method for distribution network expansion planning based on bilinear Benders decomposition is proposed. A two-stage stochastic mixed integer second-order cone programming model is constructed by stages at the planning investment level and operation optimization level. In order to avoid the high investment cost caused by extreme scenarios, the traditional Benders decomposition method is further extended, and a chance-constrained optimization method based on bilinear Benders decomposition is designed. The strong robustness performance of the proposed method is verified by three examples.