Regional modelling of soil organic carbon (SOC) dynamics is important for predicting large-scale patterns in carbon cycling and for assessing potential responses of soil carbon pools to land-use change. However, data uncertainties related to both spatial heterogeneity and small-scale differences in farming practice related to cropping systems affect the accuracy of regional models. A case study is presented from Quzhou County in the North China Plain, an area characterized by highly intensive farming. For this county, the DNDC model was validated using sampling data from 68 sites around the county under generalized farm practices. Unique modelling units based on soil type, soil texture and crop type were created and then used to model the spatial change of SOC under different farming practices. Considering the main factors affecting SOC sequestration, the results indicate that the DNDC model delivers acceptable modelling results at county level. The results show that there is a great potential for SOC sequestration in Quzhou County in its central, southern and eastern parts. Changes in farming practices show a strong effect on carbon sequestration. A very efficient and environmental friendly sequestration of SOC pools could be achieved even by decreasing nitrogen fertilizer inputs, when the amount of straw returned to the field is greatly increased.