As the proportion of distributed generation connected to the distribution network continues to increase, in order to improve the voltage stability and economy of the system, this paper proposes a planning method for distributed generation based on the improved whale optimization algorithm (WOA). Firstly, the Latin hypercube sampling and the improved K-means clustering algorithm are used to deal with the uncertainties of wind, light and load. Secondly, a load-weighted voltage stability index is proposed to quantify the network voltage stability, and then, combining with the annual comprehensive cost, a distributed power generation multi-objective planning model is established. Finally, in view of the shortcomings of the existing WOA algorithm in solving complex planning problems, the logarithmic weight distance control factor and the Nelder-Mead method are introduced to accelerate the convergence speed, and the Pareto archive evolution strategy is integrated to increase the diversity of the population. The opposition-based learning strategy is used in the searching process to prevent stuck into local minima. The simulation analysis on the IEEE 33-node system shows the effectiveness and feasibility of the proposed model and algorithm.