After almost two decades of intense research and development for mechanistic-empirical design, the AASHTO-supported Mechanistic-Empirical Pavement Design Guide (MEPDG) was finally endorsed as an interim guide for trial uses. However, local calibration of the distress models in the design guide is considered to be an essential exercise for any transportation agency before it formally adopts the MEPDG for practical design use. This paper presents the results from Ontario's local calibration of the permanent deformation models in the MEPDG. The study focused on reconstructed and new flexible pavements using field evaluation data in the pavement management system (PMS) maintained by Ontario's Ministry of Transportation (MTO). A unique feature of the study was the longitudinal calibration; that is, the local calibration parameters in the three permanent deformation models were adjusted to predict the field observed rut depths over the whole life span of a pavement section. To avoid multiple local optima of residual sum of squares (RSS) in the local calibration process, constant proportions of ruts in different structural layers were assumed to obtain the layer rut depths. A macro-based automatic procedure was developed for the local calibration. A comparison of the longitudinal calibration and pooled local calibration demonstrated the importance of the longitudinal calibration in the quantification of uncertainties involved in local calibration and the significance to pavement reliability.