In this paper, a comparison of  Bayesian estimators using non-informative priors under different loss functions, assuming  that both the stress and the strength are independently indentically burr XII random variables. Bayes estimates of  R based on extreme ranked set sampling are developed using Jeffery prior  under symmetric and asymmetric loss functions and compared with the known estimators using simple random sampling technique the Bayes estimator cannot be obtained in explicit form, and therefore extensive numerical investigation will be carried out to compare the Bayesian estimators under simple random sampling extreme ranked set sampling techniques