This paper presents an analysis of the stability and quality of the distributed generation planning problem's investment solution. A two-stage stochastic programming model is used to find the optimal distributed generators' installed capacities. We emphasize the design of scenarios to represent the stochasticity of power production from renewable sources. For scenario generation, a method is proposed based on the clustering of real measurements of meteorological variables. The quality and stability of the optimal investment solutions are thoroughly analysed as a function of the number of selected scenarios. The results show that a reduced selection of scenarios can give an inadequate solution to distributed generators' investment strategy.