Abstract: To quantify the contribution of soil heavy metal sources over a large area, surface soil in the middle and upper reaches of the Le'an River in Jiangxi Province was selected as the research object. The chemical compositions of the soil samples were analyzed to determine the local source profiles. The PMF (positive matrix factorization) model was applied to apportion the potential sources and their contribution of soil heavy metals. The results showed that the average concentrations of As, Hg, Cd, Cr, Zn, Cu, Mn, Pb in the soil samples were 21.400, 0.105, 0.25, 73.5, 88.4, 56.2, 577.0, 49.5 mg/kg respectively, ranking from 1.0 to 1.8 times of their respective soil background values of Jiangxi Province. The mean values of As, Cd, Cu, Pb exceeded the grade one of Environmental Quality Standard for Soils (GB 15618-1995) in China. The results also show that the soil heavy metals were mainly affected by lead-zinc ore smelting, gold ore smelting, copper mining, and natural sources. Lead-zinc ore smelting activities affected the soil in most areas on the left side of Jishui River Basin, with the source contribution of more than 30%. While, gold ore smelting activities currently affected only the soil near the mining area, with the maximum source contribution of 94%. In addition, the soil heavy metals around the Dexing Copper Mine was affected by copper mining and smelting activities, which the source contribution was between 19% and 89%. This study indicates that constructing the local source profiles can assist in the apportionment and verification of the validity of PMF source profiles. Combining PMF model and geostatistical spatial analysis is an effective method for obtaining the potential sources distribution, and can provide an effective support for soil pollution control decision-making.