Predicting railway track degradation is essential for ensuring safety, optimizing maintenance strategies, and reducing operation costs. Climate change has a particularly negative impact on regions affected by the phenomenon. This study develops a machine learning-based framework to predict track degradation, considering both environmental and infrastructure-related factors. Nine machine learning models were trained and evaluated on a dataset comprising track geometry, operational data, and environmental conditions from Iran’s eastern railway network. The results indicate that Random Forest outperforms other models, achieving the highest predictive accuracy. The findings reveal that increasing sand coverage leads to rapid track deterioration. This emphasizes the need for adaptive maintenance strategies to mitigate climate change effects. In order to ensure long-term operational reliability, it is essential to integrate climate resilience into railway infrastructure planning.