The implementation of a multi-microgrid (MMG) system with multiple renewable energy sources enables the facilitation of electricity trading. To tackle the energy management problem of a MMG system, which consists of multiple renewable energy microgrids belonging to different operating entities, this paper proposes a MMG collaborative optimization scheduling model based on a multi-agent centralized training distributed execution framework. To enhance the generalization ability of dealing with various uncertainties, we also propose an improved multi-agent soft actor-critic (MASAC) algorithm, which facilitates en-ergy transactions between multi-agents in MMG, and employs automated machine learning (AutoML) to optimize the MASAC hyperparameters to further improve the generalization of deep reinforcement learning (DRL). The test results demonstrate that the proposed method successfully achieves power complementarity between different entities, and reduces the MMG system operating cost. Additionally, the proposal significantly outperforms other state-of-the-art reinforcement learning algorithms with better economy and higher calculation efficiency.