Forest fires are a very dangerous problem; this disaster affects social and economic life. Detection systems are still being developed to better detect fires in the early stage. Artificial intelligence is widely used in this area, thanks to the efficiency of machine learning and deep learning algorithms for Classification, Prediction, and Detection of Forest Fires. In this paper, we will present the supervised learning methods of machine learning and stacking learning technique for the prediction of forest fires using the MODIS satellite images dataset. The objective is to study and compare the supervised learning models with the stacking learning approach using Artificial Neural Network, Random Forest, Decision Tree, K-Nearest Neighbors, Support Vector Machine, and Logistic Regression. Finally, we will resume the results of this study and expose the metrics used for the evaluation of the algorithms.