A strategy to estimate the power demand increment for Plug-in Electric Vehicles (PEV) penetration in residential customers is proposed. This is based on a stochastic model of arrivals and leavings of vehicles and has as goal to generate diversified load curves for every hour of a day, and so on, to estimate the power demand by Electric Vehicles. So, the proposed model is useful to be employed for planning purposes in distribution power systems. This study is divided in two cases, one of them has a limited number of chargers to be installed at residential customers and the second option has a free amount of charges related to number of Electric Vehicles; these represent different types of housing. This tool provides information for planning in distribution systems, based on specific information on mobility in the city to study.