This article devotes to studying the variance change-points problem in student t linear regression models. By exploiting the equivalence of the student t distribution and an appropriate scale mixture of normal distributions, a Bayesian approach combined with Gibbs sampling is developed to detect...