Malignant melanoma, which is another name for melanoma, is distinguished by variations in skin tones brought on by aberrant pigment-producing cell activity. The aim of this work is to raise the accuracy of melanoma diagnosis using dermoscopic images by employing Artificial Intelligence (AI). Pre-processing, segmentation, feature extraction, and classification are the four major phases of the predicted skin cancer detection model. Image enhancement and hair removal also seem to be part of the primary pre-processing phase of dermoscopic images. To reduce the noise in dermoscopic images, Gaussian filter is used. The K-means algorithm Clustering is the segmentation process used in this research. Regarding pre-processing, an efficient region-growing algorithm coordinates lesion segmentation. Ant Colony Optimization (ACO) is used in feature extraction. During colour morphology, morphological transformation, local features are all mined during the stage of feature extraction. Furthermore, in the classification phase, novel AI-based method is used. In order to reduce process complexity, the features are exposed to a new adapted form of the ant colony optimization process. As a consequence, the procedure is more precise and reliable. To evaluate the effectiveness of the recommended approach, the proposed method is contrasted with the existing methods.