This chapter is concerned with the SVM that defines dual parallel linear boundaries among the classes. We study the Perceptron that is a typical linear classifier and the basis for deriving the SVM. In the main part, we cover the classification process, the constraints, and the learning process of the SVM. We survey some variants of the SVM that are expansions of the standard SVM. The SVM is applicable to a nonlinear classification problem, robustly, as the most popular machine learning algorithm.