A multi-agent-based energy-coordination control system (MA-ECCS) is designed for grid-connected large-scale wind photovoltaic energy storage power-generation units (WPS-PGUs) to address the challenges of low operation efficiency, poor stability, and complex decision making. The proposed system adopts the negotiation model of the contract net protocol with the non-fixed client server cooperative mechanism among agents. When the power supply and demand are imbalanced, the system selects the state-changeable agent as the task initiator according to the load balance principle. In the energy-coordination process, the proposed system solves the global optimal-energy distribution plan of the system by considering the self-constraint and control objective of each agent and by using the improved particle swarm algorithm to achieve the maximum economic benefit on the basis of stable operations. The paper presents the overall design idea of the system, the agent models of different equipment, and the global energy-coordination optimization algorithm. An example is given to discuss the behavior characteristics of the agent and communication negotiation process and to verify that the system designed in this paper can realize the energy-coordination dispatching flexibly and efficiently. The method adopted by the system can improve the overall operation efficiency and economic benefits of WPS-PGUs. (C) 2014 Elsevier Ltd. All rights reserved.