Abstract: In this study was applied the lifetime distribution which the Lomax distribution with heavy-tail probability and the Minimax distribution which the special form of the beta distribution follow. Using this lifetime distribution, reliability model based on the non-homogeneous Poisson process was analyzed. As a result, the mean squared error, mean value function and intensity function in order to recognize the failure type for software reliability were estimated. The results of this study show that the mean squared error of the Lomax Model is smaller than that of the Minimax distribution model and the mean value function trend is smaller in terms of the true value. Therefore, the Lomax distribution Model can be regarded as an efficient model than the Minimax distribution Model. In the form of the reliability function, it gradually appears as a non-incremental pattern as the mission time elapses and the Lomax distribution Model is higher than the Minimax distribution model. Through this study, software operators will be able to use the mean square error, the mean value and the change in the intensity function to identify the type of failure in software reliability that reflects the characteristics of various lifetime distributions.