Heavy metals resulting from human activities pose significant threats to human health and the soil ecosystem. In the current study, 917 soil samples from Chongming Island in Shanghai, China, were examined for eight heavy metals. The sources of contamination were identified by using a Positive Matrix Factorization (PMF) model. Meanwhile, spatial interpolation and Moran's I index were applied to validate the model in terms of spatial linkages. The results revealed that the average concentrations of As, Cd, Hg, Pb, Cr, Cu, Zn, and Ni in the soil were 8.87, 0.19, 0.06, 28.75, 76.01, 37.74, 88.93, and 30.33 mg kg-1, respectively. The PMF analysis proved that heavy metals in the soil of the study area are mainly influenced by traffic sources (Cr and Pb), industrial sources (Zn, Cd, and Cu), station sources (Hg), and natural sources (As and Ni), with contribution rates of 22.23, 26.25, 36.38, and 15.14%, respectively. The combination of Moran's index and the spatial analysis method not only verified the analytical results of the receptor model on the one hand but also served as a supplementary explanation for the sources of heavy metals in the soil. The health risk assessment indicated that noncarcinogenic values were below the threshold values. The total carcinogenic risk (R T) of different heavy metals has a descending order of Cr > As > Ni > Cd. The R T values of multiple heavy metals for children and adults were 5.28 × 10-04 and 4.10 × 10-05, respectively, which were close to the risk threshold. Therefore, attention should be paid to the health risks, especially for children's skin contact, which is the main exposure pathway.