Public transportation systems play a vital role in urban mobility, but delays pose significant challenges, impacting passenger satisfaction and trust. This study addresses delay prediction using General Transit Feed Specification (GTFS) data comprising static and real-time information. We explore five machine learning (ML) models’ effectiveness, including Gradient Boosting, Random Forest, Support Vector Machines (SVM), Neural Networks, and kNearest Neighbors (kNN). We discuss issues such as data complexity, limitations, and model interpretability in delay prediction. Our comparative analysis evaluates these models based on predictive accuracy. SVM is consistently accurate with the lowest Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), while Neural Networks and Gradient Boosting show strong performance. Random Forest and kNN exhibit limitations. This research emphasizes the importance of accurate delay prediction and interpretable models for transportation management. The findings aid stakeholders in selecting suitable methods, contributing to improved service quality and increased public trust in transportation systems.