Gas turbines are critical assets in gas compression and power generation systems, and their reliability directly determines operational continuity and safety. This study presents a Bayesian–Weibull framework for reliability assessment of an MS5002C industrial gas turbine based on real operational failure data (N=246). As a prerequisite for prior elicitation, four candidate statistical distributions — exponential, two-parameter Weibull, lognormal, and log-logistic — were systematically evaluated and fitted using Maximum Likelihood Estimation (MLE), with model adequacy assessed through quantitative goodness-of-fit criteria including the Kolmogorov–Smirnov (K-S) statistic, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). The lognormal distribution was identified as the most suitable baseline model, yielding an initial mean time between failures (MTBF) of 123.23 days. To address data scarcity and parameter uncertainty inherent in industrial failure datasets, a Bayesian–Weibull model with an informative lognormal prior was developed and fitted using Markov chain Monte Carlo (MCMC) sampling. The posterior estimates indicated an MTBF of 115.6 days and a shape parameter of β=0.8621, indicating a decreasing hazard rate consistent with early-life or improvement-driven failure behavior. Compared with conventional frequentist methods, the proposed framework produces narrower posterior credible intervals and more stable parameter estimates, particularly under limited failure-data conditions, thereby supporting more informed maintenance decision-making. The robustness of the posterior estimates was further confirmed through sensitivity analysis with respect to prior assumptions and model parameters. Overall, the proposed Bayesian–Weibull framework provides a transparent, computationally efficient, and practically applicable tool for reliability analysis and maintenance planning in industrial gas compression systems, with potential extensibility to other turbine-based applications.