<p>This study proposes an novel optimal selective maintenance plan for mission-oriented systems, such as aircraft, ships, and power plants, where maintenance is constrained to brief intervals between missions. The primary objective is to maximize the system's remaining useful life (RUL) while minimizing maintenance costs, subject to time and resource limitations. A bi-objective optimization model is developed, incorporating three stochastic degradation processes—Wiener for non-monotonic degradation, Gamma for monotonic patterns, and Inverse Gaussian for flexible variance-mean mixtures—to estimate RUL based on real degradation data rather than lifetime distributions. The nonlinear model is linearized into a mixed-integer linear programming (MILP) formulation using auxiliary variables and Big-M constraints, solvable via tools like MATLAB or metaheuristics such as genetic algorithms. Key contributions include the pioneering integration of multiple degradation models in selective maintenance, enabling robust predictions for diverse failure mechanisms, and a data-driven approach that enhances applicability to real-world scenarios. A numerical example of a passenger boat illustrates the model's efficacy, demonstrating trade-offs between cost and reliability. This framework advances maintenance optimization, paving the way for future extensions in predictive strategies.</p>