This article presents an approach for the reliability analysis of discrete event systems, aiming to address the challenge of combinatorial explosion of states. The proposed method utilizes stochastic Petri nets and introduces an estimation technique based on average marking and throughput, which yields relevant results. This approach specifically tackles the limitations encountered by Markov models in systems with interdependent components. The introduced stochastic estimator demonstrates a convergence behavior similar to that of the Markov model in the steady state regime while eliminating the need to determine the marking graph. However, it is worth noting that the marking and throughput estimation convergence may be slower. This approach improves the reliability analysis of complex systems by taking interdependencies, by overcoming the limitations of traditional models, it offers improved prospects for reliability analysis in various fields.