Statistical models are used to understand the mechanism by which data is generated. In most models, the random variables Y_{i}, i=1,...,n, are assumed to be random samples of a distribution F, where F belongs to a class of parametric distributions. However, in many practical problems, a parametric model cannot be expected to be suitable for describing the data. In these situations, the parametric assumption can be dropped and more flexible and robust models can be used to analyze the data. In the framework of the nonparametric Bayes method, this flexibility is achieved by defining a prior distribution over the entire space of probability distributions and assuming it for the distribution of the random variable. In other words, stochastic processes are defined over a family of distribution functions and used as a prior for the random distribution. Among the most important of these priors is the Dirichlet process, which has important and interesting properties, and is therefore used in a wide range of nonparametric Bayes problems. In this article, this process and its properties are introduced.