In this paper, the geometric process model is used for analyzing constant stress accelerated life testing. The generalized half logistic lifetime distribution is considered under progressive type-II censoring. Statistical inference is developed on the basis of maximum likelihood approach for estimating the unknown parameters and getting both the asymptotic and bootstrap confidence intervals. Besides, the predictive values of the reliability function under usual conditions are found. Moreover, the method of finding the optimal value of the ratio of the geometric process is presented. Finally, a simulation study is presented to illustrate the proposed procedures and to evaluate the performance of the geometric process model.