In view of the merits and promising applications of microgrid community (MGC), this paper presents a two-level hierarchical optimization method for MGC's energy management system in smart grid environment. The lower level focuses on an individual microgrid (MG), and the upper level is responsible for managing the MGs and MG community level devices (MCLDs). It is executed consecutively from lower level to upper level in order to minimize operational costs. First, the lower level optimizes the power output of MGs generators, the charge/discharge power of energy storage system, and the exchanged power with upstream grid. The results of the lower level are transferred to the upper level as its constraints. Second, the upper level makes decisions about MCLDs' powers and the exchanged power between distribution networks based on results from the lower level while ignoring any internal constituents of MG. The model is built in ways that nonlinearity is removed by linearization approximation and nonconvexity is removed by adding implicit logic constraints. A case study based on a real system in South China's Guangxi Province is analyzed. The results demonstrate the effectiveness of the method. Of the four scenarios tested, two are proved cost-effective and are recommended for MGC operation.