<p>In this paper, we propose an improved version of osprey optimization algorithm that takes inspiration from the hunting strategy of ospreys. The original osprey optimization algorithm (OOA), often stagnates to a local optima and slow convergence speed. The concept of levy flight is used to improve the algorithm exploitation ability called levy osprey optimization algorithm (LOOA) to overcome the drawbacks of the original algorithm. The analysis of the performance of LOOA is done by its application to an industrial problem, specifically to a stainless steel utensil manufacturing unit. Markov approach is used to carry out the mathematical modeling of the system and first-order differential equations are formed in association with the state transition diagram. It is been assumed that the failure and repair rate parameters of the subsystems are exponentially distributed. Firstly, a sensitivity analysis is done using the Markovian approach to find the critical subsystem. Then OOA, LOOA and three other new metaheuristic algorithms are employed to optimize the steady-state fuzzy availability of the industrial system. The improved LOOA significantly improves solution accuracy, stability, and convergence speed, according to the results and in achieving optimum fuzzy availability in comparison to all the algorithms considered. The statistical analysis supports the better performance of LOOA. The effect of various parameter variations on the system’s availability is recorded in addition, consistently showing that LOOA outperforms OOA in attaining the maximum fuzzy availability. </p>