Abstract: During step-stress accelerated life testing (SSALT), in order to evaluate life and reliability characteristics of repairable product, the equation principle of accumulative hazard functions under two different stress levels was put forward. On the base of the rule, time conversion function was presented. After failure data were converted to the same stress condition by the way of iterative algorithm, the three-parameter Weibull process that includes the position element with Peck accelerated model was utilized to fit them. This model assumed "minimal repairing" such that any repair had no impact on the failure rate. By the use of maximum likelihood estimation, life and reliability of repairable system under normal stress conditions could be extrapolated. Residual distribution sampling was used to generate the two simulation examples. Results show that the model and method proposed is not only good at parameter evaluation with high accuracy in comparison with the traditional,but also can be applied to reliability growth testing.