The paper introduces multi-agent system (MAS) for intelligent scheduling model of multi-MG autonomy-cooperative operation. Firstly, wind power plants (WPPs), photovoltaic generators (PVs), conventional gas turbines, energy storage systems (ESSs) and controllable loads (CLs) are integrated into MGs with the price-based demand response (PBDR). Then, a 3-layer coordinate control system framework is designed for MGs, namely, a distribution management system (DMS), an MG central controller and MG controllable elements. Further, the influence of ESS modes on MG operation is discussed, namely, the longest life cycle (LLC) mode and the optimal economic efficient (OEE) mode. Finally, MORE MICROGRIDS project in Europe is taken as simulation platform. The results suggest: (1) the proposed strategy can be applied to the optimal scheduling of MG operation both in grid-connected and island modes. (2) Different ESS modes directly affect MGs scheduling, with the OEE mode optimizing MGs scheduling and with the LLC mode ensuring the ESS life cycle. (3) PBDR can smooth the load demand curve for promoting the output of WPPs and PVs. Note that the PBDR can stronger improve the MG operation in the ESS's LLC mode, because the output of the ESS in the LLC mode is lower compared with the OEE mode. (C) 2018 Elsevier Ltd. All rights reserved.