Among the most pressing challenges faced by the construction sector are project delays resulting from the complexity of the construction sector and its ingrained sources of delay risk interdependence. Machine learning provides a typical technique for solving such problems. This study identifies and develops machine learning methods for the analysis and prediction of project delay risk using objective data sources; hence, classification methods were utilized to build prediction models. Specifically, k-nearest neighbors, random forest, and decision tree were implemented when evaluating the model. By exploiting different machine learning algorithms, the performances of the proposed model are evaluated and show the best performing one during prediction. The k-NN algorithm returns the best accuracy rate of 97.21% with a higher value of response time meaning that the model created using the k-NN algorithm performs the best when used for predicting delay risks in construction projects.