With an eye on the broad vistas of extensive application of the combined heat and power（CHP） system for its environment friendliness and energy-saving， an investigation has been made on the operation optimization of a CHP system composed of PV batteries， storage batteries， heat storage tank， microturbine（MT）， heat recovery boiler， and thermo-electric loads. The characteristics of MTs electric power generation efficiency， exhaust heat recovery efficiency and pollutant gas emission are comprehensively considered. The kinetic battery model（KiBaM） is used to consider the minimum state of charge（SOC）. Finally a multi-objective chance constrained programming（CCP） model is developed with PV and stochastic thermo-electric load， operation cost， CO emission and minimum NOxemission all taken into consideration. The stochastic simulation and Pareto set are used to solve the CCP problem. A revised multi-objective particle swarm optimization（MOPSO） based on local and global memory is proposed to solve the model. Simulation results show that， with this method， suggestions can be made to optimize microgrid configuration， realize micro-sources dynamic dispatching and effectively reduce the operation cost and pollution gas emission.