Slope stability is a complex phenomenon influenced by geotechnical variables and external load conditions. In recent years, numerous slope failures worldwide have underscored the significant impact of rainfall. This has spurred increased slope stability studies to predict potential landslides or establish various scenarios. Novel computational techniques have played a pivotal role in advancing more sophisticated studies. This research proposes the integration of hybrid numerical models and the implementation of three machine learning algorithms (decision trees, artificial neural networks, and artificial neural networks with principal component analysis) to predict safety in homogeneous earth dams subjected to varying rainfall intensities. Three fundamental geometries and 40 clayey soils constituting the embankment are analyzed. All machine learnings algorithms were applied to a comprehensive model with 13 input variables and to a simplified model with six input variables. Results demonstrate a high correlation between numerical models and observations from the three implemented applications. Decision trees trained to predict the occurrence of failures achieved effectiveness exceeding 0.90 and errors of 0.08 and 0.098 in each case are reported, combined with Cohen's Kappa coefficients exceeding 0.7. Artificial neural network models trained to predict the safety factor also yielded effectiveness above 0.90 and errors below 0.13, similar to those with principal component analysis. Decision trees have the added advantage of facilitating model interpretation.