Forest fires are one of the disasters that cause problems for the environment. Forest fires can cause damage and threats, not only to forest resources but also to the entire ecosystem, both fauna and plants that can damage biodiversity and the environment of an area and can endanger human life. The source of forest fires was initially thought to come from a dry and hot environment, but in some cases, forest fires are triggered by human activities in clearing land for agriculture or other purposes. One of the factors that influence the spread of forest fires is several variables combined with humidity levels, wind speed, and rainfall. In this study, researchers used machine learning algorithms KNN and Naïve Bayes to predict forest fires and compare the results of the performance levels of each method used. The results obtained indicate that the naive Bayes method has an accuracy value of 53.33% and K-NN has an accuracy value of 62.66%