Determining the number of components in a mixture distribution is a difficult and important problem. There are various methods to determine the optimal number of components in mixture distributions, and we will mention a few of them in this article. The first method, described as the greedy EM algorithm, is based on an algorithm that adds a new component to the model at each step, and this process continues until the optimal number of components in the mixture distribution is determined. The second method is based on the maximum entropy of merging in the iteration of overlapping subclasses until the result of merging these components is that the mixture distribution under consideration has one component. This method is described as the merging of mixtures, and the third method determines the number of components of the mixture distribution nonparametrically by defining indicator variables. It is worth noting that the components of the mixture distribution considered in this article are the t-normal distribution.