In recent years, mobile ad hoc networks (MANETs) have garnered attention for their adaptability in diverse scenarios, particularly in emergency situations. Despite their potential, achieving efficient and reliable communication in MANETs remains a persistent challenge due to their dynamic and decentralized nature. A reinforcement learning-based approach is incorporated in this paper for the optimization of Ad hoc On-demand Distance Vector (AODV) routing protocol associated with MANETs. We defined state space, reward function, and action space for reinforcement learning and simulations are performed in ns2. This study not only showcases the effectiveness of reinforcement learning in optimizing MANETs routing but also establishes a foundation for future research in this domain. The paper addresses critical challenges in MANETs and outlines a promising avenue for further exploration.