Qualification of the reserve requirement of power systems in the presence of load uncertainties and renewable energy resources is one of the most important challenges of system operators. Especially, existence of wind turbines with unavoidable volatility in their generated powers makes this problem more serious. In this paper, is proposed a probabilistic method for optimal determining of the spinning and non-spinning reserves. In this method, the required optimal reserve level would be determined via simultaneously optimizing the operation cost (OC) and the expected energy not supplied (EENS) cost under security constraint unit commitment. That is, OC includes the cost of both reserve market and energy markets and EENS is computed using a piecewise linear function. The demand response program is considered to provide the reserve service from the demand side. These improve the power system reliability and also reduce the OC. Besides, energy storage systems have been included as a promising approach to diminish the uncertainties of wind power and electrical load. The method is formulated in a two-stage stochastic programming framework, where the first stage represents the day-ahead market, and the second stage deals with the real-time market. Also, a multi-step algorithm has been presented to implement the proposed model. Finally, the well-known 3-buses and 24-buses test systems would be used to verify the efficiency of proposed model.