Melanoma is a very dangerous kind of skin cancer that poses serious health risks and can be infectious if not caught early. This study investigates the use of advanced Convolutional Neural Network (CNN) architectures for automated melanoma classification using the ISIC 2019 challenge dataset. The dataset was preprocessed using methods like picture masking and hair artefact removal to increase the training data’s quality and reliability. A number of cutting-edge CNN models, such as DenseNet201, MobileNet, Xception, ResNet, VGG16, VGG19, and GoogLeNet, were implemented and evaluated in detail. Performance metrics like accuracy, precision, recall, and F1-score were used to assess each model. Among them, GoogLeNet had the best classification performance, indicating that it has a lot of promise for accurate and dependable melanoma diagnosis. The study’s findings support the efficacy of CNN-based methods in dermatological diagnostics and advance the creation of trustworthy computer-aided instruments for the early diagnosis of skin cancer. The results of the study contribute to the development of reliable computer-aided tools for the early detection of skin cancer and validate the effectiveness of CNN-based techniques in dermatological diagnostics.