<p>Under the dual pressures of energy shortages and environmental challenges, the microgrid, as a distributed energy system integrating multiple energy resources, has become one of the key technologies for the efficient use of new energy and intelligent dispatching. However, achieving optimal scheduling of microgrids still faces the complex problems of multiple objectives and constraints. This paper systematically reviews the latest research progress in the optimal scheduling of microgrids, focusing on the cooperative scheduling strategy of multi-flexible resources. The study first analyzes the composition and control methods of traditional microgrids, revealing their limitations in coping with uncertainty and multi-objective optimization; it then explores the architecture of new microgrids and their intelligent scheduling techniques, and examines the latest advances in intelligent algorithms, flexible resource optimization, and multi-objective collaborative decision-making. Based on a systematic analysis of existing research, this study finds that the current microgrid optimal dispatch still suffers from challenges such as imperfect modeling of flexible resources, high computational complexity of intelligent optimization methods, and insufficient dynamic adaptability to the market environment. To this end, this paper proposes an intelligent scheduling framework based on reinforcement learning and data-driven optimization to improve the adaptability of microgrids to uncertainty and multi-objective optimization problems. The research results can provide theoretical support and practical reference for efficient scheduling and intelligent optimization of microgrids in the future.</p>