A particular kind of skin cancer that is thought to be the most harmful when it affects people is melanoma. On the other hand, if caught early, it is treatable. Melanoma, a dangerous form of cancer, is difficult to diagnose and must be excised early. Segmenting skin lesions from pictures is a crucial step in reaching this objective. Skin lesion segmentation in a Computer-aided diagnosis (CAD) system is made more difficult by variations in the lesion’s size and form. Lesion segmentation is the initial step in CAD techniques because it yields minimal error rates in the assessment of the structure, borders, and magnitude of the skin lesion. In order to reduce the amount of diagnostic error resulting from subjectivity and intricacy in visual interpretation, computerized image analysis technology needs to be developed. By combining a novel artificial gorilla troops optimization (AGTO) algorithm based on the behaviour of the gorilla troops with the Fuzzy C-means clustering (FCM) technique all of these problems can be handled. Compared to traditional algorithms, the developed method’s results provide superior lesion segmentation. The efficacy of the developed algorithm was validated by applying on the ISIC Melanoma Classification Challenge dataset.