The inherent random and intermittence of the renewable energy resources pose a huge challenge to the Microgrid (MG) energy management systems (EMS). In order to mitigate the effects of uncertainties, we propose a novel two-stage stochastic programming model for the energy scheduling optimization by considering the uncertainties in solar and wind generation, and the plug-in electric vehicles (EV). The random nature of uncertainty is characterized by a scenarios generation approach based on autoregressive moving average (ARMA) model according to probability density function of each random variable. By use of the strategy of scenarios simulation, the stochastic problem is decomposed into the deterministic equivalent problem. The firefly algorithm (FA) is used to solve the equivalent model. The effectiveness and robust of proposed stochastic energy scheduling optimization strategy for MG is valid by comparison with the simulation results of deterministic method.