Accurate prediction and proactive management of the risk of late delivery are crucial for maintaining operational efficiency and competitive advantage. Supply chain analytics is a strategic tool that enables organizations to make data-driven decisions, improve efficiency, and enhance overall supply chain performance. This paper uses machine learning algorithms to predict the risk of late delivery in supply chains. The research also highlights the important factors that impact late delivery. The data for the study was extracted from Mendley data source. Four versions of the dataset were created for the study. The data was preprocessed, and principal component analysis was applied to reduce the dimensions. Significant features were also extracted using logistic regression and random forest. Five machine learning algorithms: logistic regression, support vector machine, decision tree, random forest, and neural networks were applied to the four dataset versions. The time taken to build the model and model size was also analyzed. Model performance measures such as accuracy, sensitivity, specificity, recall, precision, F1-score, receiver operating characteristics (ROC), and Matthews correlation coefficient (MCC) were compared. The study found that decision trees and random forests performed well consistently on the four dataset versions. Managerial implications were also provided to analyze the risk of late delivery in the supply chain.