This paper presents a Nash-equilibrium-based routing algorithm designed to enhance reliability and fairness in mobile ad hoc networks (MANETs) operating under probabilistic link failures and dynamic topologies. By modeling routing as a tripartite directed graph game, the proposed method computes stable pure-strategy equilibria that account for link reliability, energy consumption, and congestion. Unlike traditional reactive or shortest-path protocols, our approach proactively balances network load while ensuring fairness across flows, even in highly asymmetric or lossy conditions. The algorithm is fully distributed, avoids routing loops by design, and achieves deterministic convergence without relying on learning-based strategies. Implementation in NS-3 confirms significant gains in delivery ratio and fairness with minimal overhead, validating the method’s practical applicability to real-world MANET scenarios.