Machine Learning algorithms are currently being studied for their applications on image classification procedures, where with the help of training and test data, remote sensed images land covers can be classified. In this study two Machine Learning methods are being tested and evaluated to analyze their performance on urban cover classification: Support Vector Machine and Random Forest. To accomplish it, three images of different time periods (1986, 2000 and 2022) from the Atlantico Department in Colombia were obtained, training and test data generated, and the algorithms were tuned and evaluated. Finally, the results from this procedure was analysed and discussed.