The large number of decision variables of long-term planning problem for improving the distribution system causes lots of complexities. The optimum allocation of Distributed Generation (DG) sources and determining their capacity in the deregulated electric market is a key way for capacity expansion of the distribution firm. In this study, a new approach is presented to determination the optimum location and capacity of DG sources in the deregulated electric market through the net present value analysis using an optimization model. This model aims to minimize the investment cost of the distribution firm, utility cost and the losses cost considering the anticipated load peak value. Considering the power market price anticipation to be indefinite, the proposed method is based on laying out the genetic algorithm and a fuzzy model to the power market price and the capacity expansion design of the distribution system during different time intervals is proposed along with two static and semi-dynamic methods. The efficiency of the proposed approach is well shown applying it to a sample network.