This proposed model addresses the inherent limitations of conventional routing algorithms in dynamic computer networks by introducing an intelligent routing paradigm employing Q-learning. Traditional algorithms often exhibit challenges adapting to dynamic network conditions, hindering overall operational efficiency. The amalgamation of established protocols with reinforcement learning mechanisms opens avenues for advancing the robustness and efficacy of communication networks. The Q-AODV protocol seamlessly integrates AODV with Q-learning, leveraged using a modified Bellman equation, resulting in superior performance manifested through enhanced throughput, decreased end-to-end delay, and improved packet delivery ratio. The model is simulated using the Network Simulator-3 (NS3) (ns-3 Consortium in ns-3: A discrete-event network simulator (Version 3.35) Software], 2021) tool, wherein we deploy the Q-AODV protocol (Perkins et al. in. Ad hoc On-Demand Distance Vector (AODV) Routing (RFC 3561). IETF, 2023) in the network layer of all the nodes in the wireless Ad Hoc network. The improved parameters are recorded and tabulated, and graphs are created to compare the effectiveness of this innovative experiment. The process is centered on integrating a Q-table, which systematically maps network conditions to prospective actions and determines Q-values. It systematically assesses the current network state at each interval, selecting the action with the highest Q-value to identify the packet’s next hop router. The agent gradually gains a proficient understanding of network dynamics through repeated learning, enhancing the routing system. This suggested methodology demonstrates the ability to transform network routing in complex, dynamic contexts. This architecture effectively improves communication in thinly dispersed universities or similar environments, providing adaptive routing suitable for delay-sensitive applications such as real-time sensor data monitoring or emergency response systems.