Abstract: Abstract：The soil is the most fundamental means of production for human survival and life. It is of great significance to identify the factors of the potential risk because of heavy metals spatial variation in typical agricultural land for the safe utilization of heavy metals contaminated agricultural land. Normal Score Transformation was used to transfer six heavy metals data into Gaussian distribution. In this study, we identified the potential risk factors of six heavy metals spatial variation in typical agricultural land of Taihu watershed by using GeoDetector Model(GDM). Factor detector results showed that the top five potential risk factors of these six heavy metals distribution included both natural factors and social-economic factors. Interaction detector results showed that interactions of potential risk factors all enhance nonlinearly. In other words, the influence of these six heavy metals distribution were an enhance nonlinearly interaction effect, which were based on potential risk factors. The specific categories of each potential risk factor could be identified by risk detector for each heavy metal. GDM results can help identify and monitor the heavy metal contaminated areas.