Due to intensification in animal husbandry combined with trace metal supplements in fodder agricultural heavy metal inputs by manure into top soils may result in an exceedance of soil sustainability criteria within a conceivable period of time. Beside site conditions like heavy metal background concentration, soil sorption behavior or leaching rates the heavy metal status in topsoils is influenced by management practices as well, e.g. amount and quality of fertiliser or kind of land use system. We developed a site specific methodology to identify agricultural land where violation of sustainability criteria in terms of heavy metals may be affected by the economically most profitable kind of land use, which is simulated by an agroeconomic land use model. This method was firstly developed at regional scale and tested in the Dill-catchment (ca. 692 km², Hesse, Germany). On the one hand we can price opportunity costs to meet soil quality concepts and develop sustainable land use options backcoupling the heavy metal accumulation model ATOMIS (Assessment Tool for Metals in Soils) (Reiher et al., 2004) to the ProLand model (Prognosis of Land use) (Weinmann et al., 2006). Up to now the sustainability in terms of heavy metal accumulation in soils could be proven under the given model assumptions for the extensively used Dill-catchment for two tested land use scenarios (former and future common agricultural policy in Europe). On the other hand we can identify the uncertainty of the model prediction for each grid cell using Monte Carlo technique combined with Latin hypercube sampling. A sensitivity analysis for areas which may be bothered by violation of sustainability criteria reveals a detailed view to the importance of parameters which control heavy metal accumulation. Results show that the uncertainty of the heavy metal background value contributes mostly to the uncertainty of total heavy metal content in the near future. With increasing time the sensitivity of background concentration is decreasing and the sensitivity of other parameters is increasing. Although the absolute values of sensitivity measures like standardized regression coefficient or Pearson's correlation coefficient may vary widely for particular parameters on different sites, it could be shown that the uncertainty of parameters of used pedotransfer functions contribute considerably to model output uncertainty whereas the uncertainty of fertilizers' heavy metal contents could be more or less neglected in extensively used agricultural systems with generally low fertilizer inputs. ATOMIS provides a simple way of comparing different sustainability criteria as well as different pedotransfer functions regarding heavy metal contents in agroecosystems. After coupling it to ProLand the ecologic and economic assessment of agri-environmental programs in terms of heavy metal inputs into top soils is possible and the development of sustainable land use options is supported.