The study of earth movements and their occurrence factors contributes positively to the reduction of damages resulting from disasters caused by landslides, falls or alluvium. The investigation of early warning systems for landslide events that is developed for this case uses main variables for the generation of movement. In this case, the main conditioning factor is the slope, and the triggering factor is the precipitation, measured through the daily precipitation history of the mountain town of San Jose de Maipo, Chile. Using fuzzy logic rules, an algorithm capable of establishing logical relationships between variables with a prediction accuracy ratio of over 88% is developed, which implies that the applied model is efficient at the time of detecting a real warning signal.