A Multi-Agent System (MAS) based energy optimization management strategy for microgrid in market operation environment is proposed in order to maximize the cost and environment benefit. Firstly, the MAS is constructed, which consists of one market operator Agent (MOA), several MG Management Agents (MMA), one Utility Grid Agent (UGA), and many DG Agents (DGA). Then the power price bidding strategy is designed in the MOA according to the generation power plan provided by all the MMAs. By the power price bidding strategy, the best trade price and bid-winning power are determined in order to maximize economic benefit. On the basis of this, the energy management strategy is designed in MMAs, by which the optimal power dispatches for all DGAs are obtained by using an improved particle swarm optimization method for minimizing the operating cost. Finally, the validity of proposed method is demonstrated by means of simulation results.