Abstract: To solve the problem of parameter estimation of the mixed Weibull distribution in fixed number truncated data, a posteriori parameters estimation is put forth for the fixed Weibull distribution and a Gibbs sampling scheme used for Bayesian parameter estimation is designed. The stochastic simulation shows that the estimation of parameters by means of Gibbs sampling algorithm is found to be effective and marked by satisfactory stability. In particular, it proves to be much better than EM algorithm in dealing with small samples