Melanoma is the most dangerous form of skin cancer and is responsible for more than 70 percent of skin cancer deaths. Melanomas develop from malignant melanocytes. Based on the years lost to cancer, melanoma would merit a higher ranking because relatively young people are affected by this malignancy. Melanoma is usually diagnosed in patients of a relatively young age; overall, the total number of patients suffering from melanoma is accumulating. The automated computerized diagnosis mechanism helps to improve the accurate analysis of skin cancer which helps the dermatologists to accelerate the diagnostic time and improve the better treatment for the patients. Detecting melanoma skin cancer using CNN involves leveraging deep learning techniques to automatically extract and learn hierarchical features from dermoscopic images. CNNs process input images through multiple convolutional layers, capturing essential patterns such as texture, shape, and color variations indicative of melanoma. Pretrained models like SVM, ResNet, U-Net and YOLO are often used to enhance accuracy through transfer learning. YOLO, SVM, CNN, ResNet, U-net and MobileNetv3 network are employs dataset for the detection of skin cancer with 96.46%, 90.35%, 91.23%, 97.87% of success rate respectively