In view of the various forms of impact on the safety， reliability， and economy of distributed generation（DG） operation in the distribution network due to the access of large-scale distributed generation to the network for combined operation， it is necessary to carry out DG planning. According to the characteristics of DG， this paper takes into account the uncertainty of part of the distribution generation power output and the expected load value represented with fuzzy variables. Based on the fuzzy expected value model， the planning model takes minimum annual investment and operation cost as the objective function and various restrictions on economical and reliable operation as constraints. When solving the model， the hybrid encoding approach in the genetic algorithm is adopted that can accelerate the evolution. An example indicates that the planning results obtained with the method proposed have stronger adaptability to a planning environment with various uncertain factors.