The cadmium content in soils in the rural environment in the Netherlands may increase towards an exceedance of quality standards due to atmospheric deposition and the use of fertilizers and animal manure. To evaluate this problem, a simple dynamic, process-oriented model SOACAS has been developed which is aimed at predicting the accumulation of heavy metals in the topsoil in a regional context. SOACAS describes the fate of a metal in one completely mixed soil compartment using a Freundlich isotherm and analytical equations to solve the mass balance. We tested if it was possible to reconstruct the soil's present cadmium contents, using independent estimates of historical cadmium loads (‘hind-cast simulation’). About 2500 recent point observations of cadmium contents in rural areas were available. Before comparison, a map was created to translate the point information obtained from the field data to areal average information (resolution 500 × 500 m2), required for comparison with results of SOACAS. A regression model in combination with a locally-weighted smoother within the framework of Generalized Additive Modelling (GAM) was used for this purpose. A realistic geographical pattern could be obtained with very few a priori assumptions. Comparison of the map obtained by the GAM and the map obtained by hind-cast simulation showed that, despite the large uncertainties about historical cadmium loadings, the current cadmium contents were only slightly underestimated by SOACAS. Moreover, the geographical pattern for the observed and simulated contents compared reasonably well. On this basis of this exercise we believe that the model can be used to predict trends of future metal contents as a function of emission/immission scenario's. Simulations showed that cadmium contents currently decrease in highly polluted areas around industrial plants in the South-Eastern part of the country, and still increase in arable land.