Skin cancer detection and classification is one of the biggest challenges faced by the medical industry. In this paper, the automation process of solving this problem is described and implemented by using the concept of Convolution Neural Networks in Deep Learning. The basic objective is to design a reliable and effective method for determination of melanoma, the deadliest skin malignancy. In addition to fine-tuning and data augmentation, we experimented on many pre-trained models like InceptionResNetV2, ResNet, MobileNet, NASNet and VGG16.