In this paper we present a Bayesian estimation and prediction of multiplicative seasonal autoregressive models under SMN errors.