<p>High-temperature corrosion in ammonia-containing environments is an important issue for materials used in ammonia-based energy systems. However, the presence of ammonia introduces complex corrosion behavior due to the coupled effects of temperature, gas composition, and alloy composition, making systematic understanding challenging. In this study, the corrosion behavior of Fe and Fe–Cr alloys was investigated under controlled atmospheres by systematically varying temperature and ammonia (NH₃) concentration, combined with a data-driven approach to capture the resulting nonlinear behavior. Pure Fe exhibited the largest mass gain in NH₃-free environments, whereas the addition of NH₃ suppressed oxidation and significantly reduced the overall corrosion rate. In contrast, Fe–Cr alloys showed a strong dependence on temperature and NH₃ concentration, with the dominant reactions shifting between oxidation and nitridation depending on Cr content. To capture this intricate corrosion behavior, Gaussian process regression (GPR) was employed to model corrosion mass gain as a function of temperature, NH₃ concentration, and alloy composition. The model successfully reproduced the nonlinear response of corrosion behavior across these variables (<i>R</i>² &gt; 0.95). Furthermore, Bayesian optimization was applied to propose promising experimental conditions and alloy compositions for improved corrosion resistance. This data-efficient framework enables accurate mapping of corrosion behavior over a wide parameter space with a limited number of experiments and provides practical guidance for material selection in high-temperature ammonia environments.</p>