The AASHTOWare Pavement ME Design software was developed as an advanced pavement design tool based on the mechanistic-empirical pavement design guide (MEPDG). The existing performance prediction models utilized in this design guide are nationally calibrated, and it is essential to calibrate these models based on the specific regional materials and environmental conditions and pavement performance information. In this study, the Pavement ME design prediction models were calibrated using local inputs and performance data for the state of Oklahoma. Also, to facilitate using Pavement ME, an interface software named INput ME was developed. This software can be used to process and compile the required pavement ME input data based on the Oklahoma material, traffic, and climate properties gathered from long term pavement performance and Oklahoma and national cooperative highway research program datasets. The material input data were evaluated, and the most accurate available data was selected. The predictions from distress and international roughness index (IRI) models were evaluated and compared with the measured distress values, and the accuracy and bias of each model were determined. The nationally calibrated models showed large errors and significant bias values, which asserts the need for local calibration. The locally calibrated coefficients for the distress and IRI models for the Oklahoma pavement system were determined. The predictions from calibrated models show that the use of calibrated coefficients improves the accuracy of the pavement ME predictions and the design of flexible pavements in Oklahoma.