In this study, after applying the Exponential-exponential and Inverse-exponential distributions to the NHPP software reliability model, the reliability performance of the applied model was newly compared and evaluated with the Exponential-basic model. For this study, the failure time data collected during software system operation was used, and the parameter estimation was solved by utilizing the maximum likelihood estimation (MLE) method. As a result, first, in the analysis of the performance pattern using the mean value function, the Exponential-exponential model with the smallest error in predicting the true value showed efficient performance. Second, in the evaluation of the intensity function, the failure occurring rate of the Exponential-exponential model showed the smallest value at the initial stage and continued to decrease with the lapse of the failure time, so it was evaluated as an efficient model. Third, as a result of analyzing the future reliability performance by applying the mission time, the Exponential-exponential model showed stable high performance, but the Inverse-exponential and the Exponential-basic model showed inefficiency in which the performance continued to decrease. In conclusion, it was found that the Exponential-exponential model has the best performance among the proposed models. Through this study, the reliability performance of distributions with exponential-type attributes were newly identified, and basic design data that could be utilized in the development process could be presented to software operators.