We　propose　a　fully　Bayesian　inference　for　semiparametric　joint　mean　and　variance　models　on　the　basis　of　B-spline　approximations　of　nonparametric　components.　An　efficient　MCMC　method　which　combines　Gibbs　sampler　and　Metropolis-Hastings　algorithm　is　suggested　for　the　inference,　and　the　methodology　is　illustrated　through　a　simulation　study　and　a　real　example.　(c)　2013　Elsevier　B.V.　All　rights　reserved.