Traffic volumes are widely believed to be key issues in the change in road surface conditions. The present study analyzed the impacts of traffic volumes of different vehicle categories on the change in road roughness as indicated by International Roughness Index (IRI) in Indonesia inter-city roads. The road volume-to-capacity ratio (VCR) was also considered in the modelling. The study used the data of 13 roads spanning 139.63 km on the north coast of Java. The analysis was made on three categories of vehicles: heavy vehicles, light vehicles and motorcycles, which were regrouped from the initially seven categories surveyed. A fixed-effects regression model for panel data was applied with repeated measurement over six years. The fixed effects method assumes that the intercept is different between subjects and the slope is the same between subjects. Using linear and non-linear regressions (Generalized Additive Model), the results show that heavy vehicles' AADT (Annual Average Daily Traffic) determines the change in IRI values. The findings will contribute to the management of road sustainability.