Supervised learning is a central topic in statistics and machine learning. Two major types of supervised learning tasks are regression and classification. There is a rich literature of methods, algorithms, and theory developed for both regression and classification, including classical results for low‐dimensional problems and modern techniques for high‐dimensional data analysis. This article provides a review on supervised learning in terms of its basic principles, decision theory, and methods and tools. Model selection is a critical issue in supervised learning, and recent results in variable selection are also discussed.