Accurate segmentation of skin lesions in dermoscopic images is essential for the early diagnosis and treatment of skin conditions, including melanoma. This study evaluates the performance of the total variation Chan–Vese model, a level set-based segmentation method, on the HAM10000 dataset, which encompasses seven diagnostic categories. By evolving contours based on intensity differences, the total variation of the Chan–Vese model provides significant advantages in segmentation. We employed pre-processing steps such as noise removal, image enhancement, and artifact elimination to enhance accuracy. The segmentation performance was assessed using the Jaccard Index, revealing high accuracy across various lesion types and demonstrating the model’s adaptability and robustness. Despite encountering computational limitations, this study underscores the potential of the total variation Chan–Vese model in clinical skin lesion analysis.