In this study, attributes of software development cost were evaluated by applying the exponential type distributions (Exponential-exponential, Inverse-exponential) which are utilized in the reliability testing field to the software development model. Also, the proposed distribution models were compared with the GoelOkumoto basic model to verify cost property, and the optimal development cost model was presented. For this study, a total solution was performed using software failure time data generated during desktop application operation, parameter calculations were solved using the maximum likelihood estimation (MLE) method. As a result, fist, when the testing cost per unit time and the cost of eliminating a single fault detected during the development testing process increase, the development cost increases, but the release time does not change. But, if the fault correction cost detected by the operator during normal system operation increases, the development cost increased along with the delay of the release time. Second, it can be confirmed that the Exponential-exponential distribution model is the most efficient among the proposed models as it has the best performance in terms of development cost and releasing time. Third, if software developers and operators can utilize this analysis information efficiently, they can predict and design a reasonable development process by analyzing the related cost and time attributes.