This paper explores a machine learning model designed to predict permissible tilting angles for tilting trains, aiming to enhance travel efficiency and passenger comfort. The model utilizes an interpolation algorithm to forecast the tilting angle based on two key inputs: the speed of the train and the radius of curvature of the track. Historical data on train speeds, track curvatures, and corresponding tilting angles are collected and preprocessed to develop a robust dataset. Polynomial feature engineering is employed to capture the nonlinear relationships between inputs. A linear regression model is developed on these features to accurately predict tilting angles. The model’s predictions enable dynamic adjustment of the train’s tilting mechanism, ensuring optimal performance on curved tracks. This approach offers a cost-effective solution to improve train speed and stability on existing rail networks without extensive infrastructure modifications. The proposed model is evaluated using real-world data, demonstrating its effectiveness in providing smooth and safe travel experiences. The incorporation of this model into the train control systems promises significant improvements in rail transport efficiency, passenger comfort, and operational reliability. This research underscores the potential of machine learning in modernizing and optimizing rail transport systems.