In this paper, field measurements were performed to determine SO 2 concentration gradients from a highway in Shanghai using passive samplers. It was demonstrated that passive sampling method is a cost-effective and convenient way to monitor specific gaseous pollutants at small scales over long sampling periods in air quality studies. Using function fit analysis for the measured results, a shifted power–law relationship had been found between SO 2 concentration and the distance from a highway. Accordingly, an empirical shifted power–law model was developed for describing and predicting the SO 2 gradients near a highway, in which k is the only parameter and named as diffusion attenuation coefficient. There was a surprisingly significant negative correlation between known SO 2 concentration at reference point (C 0 ) and diffusion attenuation coefficient (k). By the correlation equation of C 0 and k, appropriate value of k could be calculated with measured C 0 . Therefore, the empirical shifted power–law model developed in this study could be practically and conveniently applied for predicting the SO 2 distributions near a highway with known C 0 .