Soil heavy metal contaminated sites with multiple sources of pollution have caused worldwide public concern. However, the lack of correlation of risk assessment between source identification of heavy metal led to unclear direction of source governance. A methodology was established by combining source apportionment of human health risks with ecological enrichment to characterize source-identified risks of heavy metals based on Ordinary kriging interplotation. Principal component analysis (PCA) and positive matrix factorization (PMF) model were used to identify and classify potential sources of heavy metals synthetically. The integrated results were incorporated into the health risk model to evaluate potential non-carcinogenic and carcinogenic risk of soil heavy metals. A case study was conducted in Suzhou city of China. The results showed that concentrations of Cd and Hg were highly above the background values, accounting for percentages of 239.6% and 415.9% above background values, respectively. The source contributed human health risk index of As contributed 76.9% to non-carcinogenic risk by pollutant sources of agriculture activities. The Non-cancer health risk index for children and adults was 1.08 and 1.00 respectively. The cancer health risk was 3.67E-03 for children and 3.97E-04 for adults. Cr originated from indutriy activities, accounting for 29.5% of total heavy metals, and constituted the largest carcinogenic impact on the population. This study provided a new insight for the treatment of mutil-sources of soil heavy metal pollution and also some reference value for the improvement of the risk assessment system.