When the Verhuslt grey model is used to predict landslides, errors occur sometimes. In this paper, the fundamental principles of the Verhulst model are presented and the causes for such prediction errors are analyzed theoretically. The conditions for the application of the Verhulst model for landslide prediction are proposed. A Case study shows that the Verhulst model is remarkably reliable for prediction of short period landslides.