Quantifying source-oriented risk can identify primary pollution sources to help alleviate risks to ecosystems and human health posed by soil heavy metals. Taking Yulin National Energy and Chemical Industry Base as an example, ecosystem and human health risk assessments of each identified source category were quantitatively calculated by combining the Potential Ecological Risk (RI) and Total Carcinogenic Risk (Total-CR)/Total Hazard Index (Total-HI) assessment models with the positive matrix factorization (PMF) receptor model. In this work, an analysis of the Geoaccumulation Index (Igeo), Contamination Index (Pi), RI, CR and HI of heavy metals (As, Cr, Cu, Zn, Cd, Pb, Hg and Mn) identified universal ecosystem risks and both carcinogenic and noncarcinogenic health risks in most sites. Source apportionment results indicated that the dominant source of heavy metals in the soil was coal-related activities (52.5%), followed by industrial activities (22.0%), traffic activities (13.2%) and agricultural activities (12.3%). The source-oriented quantitative risk assessment results showed that coal-related activities are the greatest contributor to RI and Total-HI, while industrial activities are the largest source of Total-CR, which should be controlled, to reduce the carcinogenic health risk posed by exposure to heavy metals. Display omitted •The level of heavy metals in soils from Yulin were investigated.•Igeo, Pi, RI, CR and HI values in soil indicated high risk to ecological and human health.•Four source categories were identified using PMF model.•Source contributions to ecological and human health risks were quantitatively and spatially identified.•Coal-related activities caused the greatest potential ecological risk and hazard index.