With the growing impact of global warming and depletion of fossil fuels, the necessity of renewable energy resources is becoming more and more obvious. However, investing in intermittent renewable resources, such as wind power, by only considering the regional power potentials jeopardizes system reliability and increases the necessity of spinning reserves which balance the instability in power systems. In this paper, we suggest a stochastic mixed integer nonlinear programming model that decide the regional allocation of wind power capacity with the objective of minimizing the needed spinning reserves to compensate for the negative effects of intermittent nature of wind power. We also suggest a heuristic approach for solving the suggested model. Both the stochastic mixed integer nonlinear programming model and the heuristic approach are new in the literature. Parameters used for numerical analysis are gathered from real-life data sets which are provided by Turkish State Meteorological Service, Energy Exchange Istanbul (EXIST), Turkish Electricity Transmission Corporation and General Directorate of Renewable Energy. The regional distribution of wind power capacity obtained from our approach is compared with the existing installed wind power distribution.