This paper obtains a new estimates of the parameters of the generalized life testing model by using maximum likelihood, Bayesian and Lindley Bayes approximation procedures. The risks of these estimates were compared to the risks of moment estimates (EL-Shahat and Shalaby (1995)) and structural estimates (Hoq et al (1974)) based on a Monte Carlo simulation study. A class of quasi-prior distribution was used in the comparison for both Bayes procedures. The Bayes estimate of the reliability function of the generalized life model is also obtained by using the Bayes approximation form due to Lindley (1980). The estimated risks of the approximate Bayes estimates of reliability function are compared with the corresponding estimated risks of the maximum likelihood estimates (MLE's).