Mobile Ad Hoc Networks (MANETs) function in extremely dynamic and energy-constrained settings where extended network lifetime depends on effective routing. Reactive and proactive methods are among the traditional routing protocols’ problems, which include significant control overhead, excessive energy consumption, and limited scalability. Using Q-learning to dynamically modify routing decisions based on energy prediction and hybrid chain-clustering techniques, we present a novel Reinforcement Learning-Based Energy-Aware Hybrid Routing (RL-EAHR) protocol to address these problems. RL-EAHR greatly lowers control overhead, increases energy efficiency, and boosts overall network adaptability by combining reinforcement learning with energy-aware routing. According to simulation results, RL-EAHR outperforms traditional MANET routing protocols in terms of packet delivery ratio, energy usage, and network lifetime, indicating which it is a viable option for scalable and energy-efficient MANET communications.