Microgrid could allow renewable and clean resources to penetrate into a controllable utility and achieve maximum utilization of existing energy and demand-side management. An optimization model of economic operated CHP microgrid system considering heating income is established. The microsources can provide both active and reactive power in the model. A typical microgrid consisting of wind turbine, photo voltaic, storage battery, micro turbine, fuel cell, thermal and electric loads is selected as the research object. Based on the conventional economic dispatch strategies of microgrid, improved dispatch strategies are proposed and comparative analysis of their impact on the optimization results is conducted. Improved genetic algorithm is chosen for the optimization of this system and comprehensive cost of one day in grid-connected mode considering spot price and islanded mode assuring power supply reliability of sensitive loads, results are comparatively analyzed in detail. So the validity of the proposed model, dispatch strategy and algorithm is proved.