The probabilistic nature of pavement performance has led to the incorporation of reliability into pavement design procedures. A robust and comprehensive reliability analysis methodology that is coupled with accurately characterized inputs variability is necessary to derive the full benefit from a reliability analysis. This article evaluates pavement input variability impact on the design reliability and variability of rutting performance. It further studies systematically the variability of inputs such as layer modulus and thickness. This was achieved using inputs and performance data from several accelerated pavement testing (APT) structures and actual field sections. The variability of inputs such as layer modulus and thickness was established using measured structural and material properties. The ERAPave performance prediction tool was used to predict the rutting performance of the APT structures, whereas Monte Carlo simulation (MCS) was employed to generate the associated dispersion of predicted rut depth. A two-component reliability analysis methodology that uses the first-order reliability method (FORM) was used to estimate the reliability of the studied sections. A discrepancy in input variability is observed between the APT and field sections, whereas the MCS generated rut depth variability exhibits a wide variation. The FORM-based two component reliability analysis methodology has delivered reliability estimates that are reasonable and acceptable.