Abstract This paper summarizes some results of beta regression models and proposes a Bayesian method to fit these models, including joint modeling of the mean and dispersion parameters. This method is implemented through simulated and applied studies.