The characteristics of integrating distributed generation (DG) in a distribution network have changed. These electrical resources are used as an alternative energy source to the main grid. The technical and economical benefits of such units are achieved only when they are optimally sized and placed in the network. In this paper, a static mixed integer non-linear model for distributed generation planning is defined and solved using a modified NSGA method (Non-dominated Sorting Genetic Algorithm). Different DG technologies are considered and the objective functions for minimization are defined as the total active loss, investment and operational costs, and environmental pollution. The method is applied to a test distribution network and the results are discussed in comparison with other methods.