Road deposited pollutants (build-up) are continuously re-distributed by external factors such as traffic and wind turbulence, influencing stormwater runoff quality. However, current stormwater quality modelling approaches do not account for the re-distribution of pollutants. This undermines the accuracy of stormwater quality predictions, constraining the design of effective stormwater treatment measures. This study, using over 1000 data points, developed a Bayesian Network modelling approach to investigate the re-distribution of pollutant build-up on urban road surfaces. BTEX, which are a group of highly toxic pollutants, was the case study pollutants. Buildup sampling was undertaken in Shenzhen, China, using a dry and wet vacuuming method. The research outcomes confirmed that the vehicle type and particle size significantly influence the re-distribution of particle bound BTEX. Compared to heavy-duty traffic in commercial areas, light-duty traffic dominates the re-distribution of particles of all size ranges. In industrial areas, heavy-duty traffic re-distributes particles > 75 mu m, and light-duty traffic re-distributes particles < 75 mu m. In residential areas, light-duty traffic re-distributes particles > 300 mu m and < 75 mu m and heavy-duty traffic re-distributes particles in the 300-150 mu m range. The study results provide important insights to improve stormwater quality modelling and the interpretation of modelling outcomes, contributing to safeguard the urban water environment.