This study investigates the occurrence and spatial distribution of potentially toxic elements (PTEs) (Hg, Cd, Cu, Mo, Pb, Zn, Ni, Co, Cr, Al, Fe, Mn, V and Sb) in 67 road dust samples collected from urban industrial areas in Ahvaz megacity, southwest of Iran. Geochemical methods, multivariate statistics, geostatistics and health risk assessment model were adopted to study the spatial pollution pattern and to identify the priority pollutants, regions of concern and sources of the studied PTEs. Also, receptor positive matrix factorization model was employed to assess pollution sources. Compared to the local background, the median enrichment factor values revealed the following order: Sb>Pb>Hg>Zn>Cu>V>Fe>Mo>Cd>Mn>Cr≈Co≈Al≈Ni. Statistical results show that a significant difference exists between concentrations of Mo, Cu, Pb, Zn, Fe, Sb, V and Hg in different regions (univariate analysis, Kruskal-Wallis test p<0.05), indicating the existence of highly contaminated spots. Integrated source identification coupled with positive matrix factorization model revealed that traffic-related emissions (43.5%) and steel industries (26.4%) were first two sources of PTEs in road dust, followed by natural sources (22.6%) and pipe and oil processing companies (7.5%). The arithmetic mean of pollution load index (PLI) values for high traffic sector (1.92) is greater than industrial (1.80) and residential areas (1.25). Also, the results show that ecological risk values for Hg and Pb in 41.8 and 9% of total dust samples are higher than 80, indicating their considerable or higher potential ecological risk. The health risk assessment model showed that ingestion of dust particles contributed more than 83% of the overall non-carcinogenic risk. For both residential and industrial scenarios, Hg and Pb had the highest risk values, whereas Mo has the lowest value.