Heavy metal pollution in the river environment has been a source of widespread interest due to potential threats to human health and ecosystem security. Many studies have looked at heavy metal pollution in the context of single source-pathway-receptor relationships, however few have sought to understand pollution from a more wholistic multi-media perspective. To investigate potential risks in a more wholistic way, concentrations of six heavy metals (Cd, Cr, Ni, Cu, Zn and Pb) were detected in multi-media (water, sediment and riparian soil) collected from 14 sampling sites in the main stream of the Songhua River. Chemical analyses indicated that the average concentration of heavy metals in water followed: Zn > Cr > Cu > Pb > Ni > Cd, with a different trend observed in sediments and riparian soil: Zn > Cr > Ni > Pb > Cu > Cd. The potential risk was evaluated using the heavy metal pollution index (HPI), Nemerow pollution index (P-N), hazard index (HI) and carcinogenic risk (CR) metrics. Results showed that all HPI values were lower than the critical level of 100 indicating that the levels of these targeted heavy metals were within drinking water safety limits. The P-N indicated that both sediment (2.64) and soil (2.95) could be considered "moderately polluted", with Cd and Zn providing the most significant contributions. A human health risk assessment suggested that the non-carcinogenic risks were within acceptable levels (HI < 1), as was the cancer risk associated with dermal adsorption (CR <10(-6)). However, the CR associated with ingestion exposure (4.58 x 10(-6)) exceeded the cancer risk threshold (10(-6)) indicative of elevated cancer incidence in exposed populations. Health-risk estimates were primarily associated with Cd in the Songhua River. Source apportionment was informed by Pearson correlation analysis coupled with principal component analysis (PCA) which indicated that Cu was mainly derived from natural (geogenic) sources; Cr and Ni were associated with industrial emissions; Pb might be derived from agricultural and transportation sources; Zn might be from industrial, agricultural activities and transportation; while Cd is likely from industrial and agricultural emissions. The source apportionment information could provide the basis for a risk-management strategy focused on reducing Cd and Zn emissions to the riverine environment. Results from this study will provide the scientific knowledge that is needed for measuring and controlling heavy metals sources and pollution characteristics, and identifying the potential cancer risk with different exposure pathways, as well as making effective environmental management policies at catchment or regional scales.