Skin cancer is an abnormality in skin cells caused by mutations in cells Deoxyribonucleic Acid (DNA). Most deaths from skin cancer are caused by the malignant type. Therefore, one of the last types of cancer is considered a treatment that can detect the disease early by biopsy examining, so the best solution for improving the diagnosis of skin cancer is early detection. Computer-Aided Diagnosis (CAD) is one of the widely used imaging techniques for detection and classification of skin cancer. The automatic detection and classification of image is considered very important for tumors skin and very challenging task for medical images. This thesis presents a proposed system for classification of skin cancer after its detection with the help of deep learning mechanisms and machine learning algorithms, where several steps are used in the form of stages, which are include, the image acquisition stage, image pre-processing, and the classification stage. The used dataset is obtained from the ISIC (International Skin Image Collaboration) Archive, it contains 3297 images. There are 1497 image cases of malignant skin caser type, and 1800 images cases for benign. In preprocessing stage, hair removal algorithm is using. The First proposed model depends on Convolutional Neural Networks (CNN) classifier. The second proposed model uses Naïve Bayes (NB) classifier. While the third proposed model relays on Support Vector Machine (SVM) classifier. And each model with applying preprocessing algorithm and without applying. The results show that the first proposed model using (CNN) without preprocessing had average accuracy 85.00%, while with preprocessing had accuracy 69.99%. The second proposed model using (NB) without preprocessing had average accuracy 70.15%, while with preprocessing had accuracy 69.69%. The third proposed model using (SVM) without preprocessing had Achieve accuracy 76.81%, while with preprocessing had accuracy 77.12 %.