<p>Change-points occur during the testing phase of software development owing to changes in the testing environment and factors related to specific software development technology. These change-points reduce the goodness-of-fit of software reliability growth models (SRGMs) to actual data, thereby degrading the reliability assessment accuracy. Several SRGMs that consider the occurrence time of a change-point as a model parameter have been proposed thus far; however, they use the change-point occurrence times in the testing environment, such as when a decision is made to implement software development management by the manager. Consequently, the change-points resulting from factors related to specific software development technology and their detection methods have not been sufficiently discussed. In this study, “Change Finder,” an anomaly detection method that statistically detects a change-point in the fault-counting data, is introduced. The detected change-points were applied to a delayed S-shaped SRGM with a change-point because the fault-counting data exhibited an S-shaped software reliability growth curve behavior. To evaluate the effect of incorporating change-points on the model accuracy, the goodness-of-fit of the proposed model was compared with that of a model considering a change-point in the testing environment. Results revealed that the model accuracy was improved when the detected change-point was used as a model parameter. In addition, the effectiveness of applying Change Finder as a change-point detection method was confirmed by comparing its performance with that of the Laplace trend test, an existing change-point detection method.</p>