When caught in its early stages, it is simply curable. Effective early detection and classification techniques are required for skin cancer, a worldwide health concern with rising prevalence. Convolutional Neural Network (CNN), a subfield of Deep Learning, is a widely used technique in a variety of applications, particularly in the medical field. The seven skin cancer types —basal cell carcinoma, dermatofibroma, melanoma, melanocytic nevus, vascular lesion, benign keratosis, and actinic keratosis—are classified using a CNN model developed for this research. With the use of pre-trained ImageNet weights as well as fine-tuning (FT) the CNNs, we trained VGG19 on a total of 10,015 images that have been collected from Human Against Machine data-set (HAM 10000), which is accessible on Kaggle. Transfer learning (TL), which improves diagnostic accuracy, has become a potent technique in this field. However, the best results were shown with FT. The developed model had an overall accuracy that was higher than that of the model employed in TL, which was 78.6 ± 2. 7019.We were able to determine precision, recall, AUC, and F1_score values for every class for TL and FT for 5 Run.