In the last decade, an extensive amount of pavement research has focused on permanent deformation characterization of pavement materials and on the development of models for predicting rut depth. The extent of the effort is an implicit recognition that rut depth formation is a serious form of pavement structural distress affecting both quality and safety. Various models for predicting rut depth based on the rational approach and on the statistical regression of field results approach are suggested. The paper deals with the statistically-based models which are found very attractive and useful for practical engineering applications, such as the Pavement Management System, at the network level. A brief review of the existing models is presented, which illustrates their limitations and drawbacks.