A state-of-the-art commercial drones have typically battery limitations to support their operation. Effective battery management is one of the main enablers of practical drone system operations. This paper designs an energy-efficient monitoring drone-based scheduling algorithm to control the number of activation drones in a multi-drone system. This study focuses on reducing overlapping regions of the inter-drone monitoring area in terms of energy efficiency. We formulate this problem as finding a maximum weight independent set (MWIS) problem on a graph model and propose a real-time update of the message algorithm based on belief propagation. Our results showed that the proposed algorithm achieves significant improvements in energy consumption compared to the existing algorithm.