This study examines a relapse stochastic epidemic system with a non-linear random transmission to enhance the model’s realism. The uniqueness of a strong positive solution is proved, and its asymptotic behavior is studied using Lyapunov methods, deriving the necessary conditions for the average extinction and persistence of the disease. An optimal control strategy is then formulated for both deterministic and stochastic cases. For a quadratic cost function, we apply the Pontryagin maximum principle in its strong form, which is more comprehensive in the general setting. Even when processes rely on accumulated information about the state’s evolution, we establish the necessary conditions and characterize the optimal strategy using the Hamiltonian function and adjoint equations. Furthermore, we analyze the implications of these results in the context of epidemic control. Finally, numerical simulations validate the theoretical results and enhance understanding of the model’s behavior under various conditions.