This work explores Support Vector Machines and Convolutional Neural Networks. In the first experiment, we apply SVM with different kernels to classify the Keel Bananas Dataset. The performance is compared with a Decision Tree Model. We achieve best results with the RBF kernel (90%) and worst with linear kernel (55%), whereas the Decision Tree Model exhibits accuracy in order of 78%. In the second experiment, CNN models are constructed to to classify the MNIST dataset. We propose seven different models, and the best one achieves accuracy of 99%. Our conclusion is that both SVM and CNN are powerful machine learning methods.