<p>Reliability assessment is an essential part of designing and maintaining industrial computing systems. This is especially true in mission-critical settings like automated manufacturing lines, aerospace systems, and chemical plant controls. In this study, we create a framework for estimating reliability in a two-module redundant industrial system facing uncertain operating conditions. We model the failure rates of system components using Triangular Fuzzy Numbers (TFNs) to account for imprecision from environmental and operational variability. We use a Markov process to describe the system's changes between working and failed states, and we formulate the relevant Kolmogorov differential equations to calculate state probabilities over time. We optimize the fuzzy failure rates with a Fuzzy Particle Swarm Optimization (FPSO) algorithm based on α-cut intervals. Our method combines fuzzy modelling, Markov reliability theory, and soft computing to efficiently calculate reliability bounds. A numerical example demonstrates the effectiveness of our approach. This research helps improve reliability prediction in safety–critical industrial applications, especially where uncertainty is involved.</p>