Melanoma is the sixth leading cause of death in the United States, with more than nine thousand people succumbing to the disease each year. The earlier melanoma is detected and treated, the longer a person can expect to live after being diagnosed with the disease more analytical developments are still necessary. The skin injury limit anomaly, which addresses the "B"component included in the "ABCD rule,"is regarded as an essential clinical component for the early detection of melanoma. In addition, the blue-white line structure evacuation is another method that helps further strengthen the capability of recognition. In this study, we provide an AI-based localization approach for recognizing skin disease boundary discrepancies. The system was developed for use in dermatological research. The process entails removing the skin disease from the dermoscopic images, identifying the skin lesion, estimating line inconsistency, preparing learning models known as SVM, RF, DT, and KNN gathering move learning to distinguish line anomaly naturally, which ultimately leads to a decision as to whether or not the skin lesion boundary is predictable or not. The approach is very favorable outcomes, with an accuracy rate of 93 percent, a sensitivity rate of 91.6 percent, a specificity rate of 92.8 percent, and an F-score of 95.4 percent, respectively.