This paper discussed a new scheme is called joint Type-II generalized progressively hybrid censoring scheme (JGPHCS-II). It assumed that the lifetime distribution of the items from the two populations follow exponential distribution. Based on the JGPHCS-II , we first consider the maximum likelihood estimators of the unknown parameters along with thier asymptotic confidence intervals. Next, we provide the Bayesian inferences of the unknown parameters under the assumptions of independent gamma priors on the scale parameters using squared error (SE) , linear-exponential (LINEX) and general entropy (GE) loss functions. Using gamma conjugate priors, the Bayes estimators are developed relative to both symmetric and asymmetric loss functions. Two numerical application based on simulated and real data sets are analyzed to discuss how the applicability of the proposed methods in real phenomenon. Finally, to examine the performance of proposed methods, a Monte Carlo simulation study, simulated example and real-life data are carried out.