Mobile ad hoc networks (MANETs) are communication systems that are both decentralized and self-organizing. They are now ubiquitous in the electronic infrastructure of today. Services requiring mobile infrastructures, constant connectivity, and adaptable protocols like military operations have found MANETs to be an important solution. This emphasizes the need for more accurate estimates of the reliability of mobile devices. Mobile ad hoc networks rely heavily on the reliability of its routing infrastructure, hence having a reliable routing system is essential. As a means of bolstering trust between routing nodes, a number of strategies such as encrypted systems, trust leadership, and authoritative routing protocols have been proposed. It is challenging to dynamically determine the suspicious behaviors of routing nodes, making it so that most routing systems are impractical in practice. Meanwhile, there is currently no reliable method for protecting networks from attacks by malicious nodes. Reinforcement learning (RL) techniques are used in this study to present a trusted routing strategy called reinforcement trust-based routing (RTR), which is designed to correctly portray the reliability of recommendations in mobile ad hoc networks. The reinforcement learning approach is used to help routing nodes make dynamic decisions about which links to employ for optimal security and performance. Despite operating in a network where half of the nodes are malevolent, our routing strategy still outperforms competing algorithms in terms of latency, as shown by our experimental results. Indicators of performance, such as energy use and throughput, corroborate the practicality and efficiency of our strategy.