In recent years, by considering problems such as environmental pollution and energy crisis, using renewable distributed generation resources as a clean energy for supplying load indistribution network is growing. On the other hand, wind energy as a free and renewable energy has been always considered. So, in this paper, by using NSGA-II multi-objective optimization algorithm, placement of wind turbines for reducing losses and improving Loadability margin and voltage profile of distribution network has been investigated. Productivity generated power of these resources on the base of environmental situation has a probabilistic nature so using probabilistic methods is essential. However, for reducing calculations and speeding up time for solving these probabilistic problems, methods which are on the base of variable data classification are used. In this paper, by using K-means classification, wind turbines data and network Load are divided into the several clusters and then network for these clusters is analyzed. Results of running this algorithm in network show fastness and accuracy of this method.