In this paper we devote ourselves to a general growth curve model with power transformation, random effects and AR(1) dependence via a Bayesian approach. Two priors are proposed and both parameter estimation and prediction of future values are considered. Some numerical results with a set of real data are also given.