Supervised learning is a type of machine learning where the algorithm is trained on a dataset that includes both input features and the correct output. It learns to map inputs to outputs and is commonly used in tasks like classification and regression.