Life-cycle management based on the degradation diagnosis is useful for preventing an unexpected failure and extending service life of electric power apparatuses, minimizing life-cycle cost. In our previous study, we formulated a numerical model of life-cycle management based on time-based maintenance and condition-based maintenance. Then, we applied the model for evaluating the economic effect of degradation diagnosis of power cables, assuming that the maximum length of water tree as a replacement criterion can be measured nondestructively. To carry out reliable life cycle management, however, accurate data on the relation between extent of degradation and failure probability or remaining life are necessary. In this study, we examine the influence of accuracy of the data used to determine the optimum diagnostic parameters and evaluate how the life-cycle cost is affected by the employment of inaccurate data. The results show that the condition-based maintenance with degradation diagnosis can be less subject to the accuracy of the back data and is possible to realize the reliable life-cycle management.