Abstarct: Bayesian method is one of the paradigms used in statistical inference , where the parameter θ is a random variable with probablity distribution function π(θ) on Θ. In this method, the posterior distribution is used as a criterion to detrmine the Bayes estimator. Due to the rapid growht of Bayesian method, since twenties century, in this thesis we will focuse on this type of estimator. In this regard, we find the generalized Bayes estimator under the linex loss as well as the reflected normal loss. Then, we show that for the normal and the scale mixture of normal distributions, under the linex loss, the generalized Bayes estimator is a soft wavelet shirinkage estimator. Furthermore, taking this importance, that admissible and minimax concepts play an important role in the statistical inference and decision making, we will investigate these estimators in terms of the concepts.