Reinforcement Learning (RL) is a promising technique for scheduling and planning storage systems in microgrids, which are small-scale power networks that can operate independently or in coordination with the main grid. RL can enhance the utilization of local renewable energy sources and reduce the operational costs of microgrids. In this comprehensive study and review the state-of-the-art applications of RL for Microgrid Energy Management (MEM), focusing on battery storage systems is discussed. This work also identify the main challenges, limitations, and future directions in this domain. Furthermore, this article present a novel benchmark algorithm that compares the performance of RL with mixed-integer linear programming (MILP), a widely used optimization technique. This complete work provides a valuable insight into the current status and future prospects of RL for MEM.