Crop model is constructed on the field scale. There are simulated errors due to the up-scaling when it is applied from filed scale to larger scale. Regional applications of crop model are becoming universal methodologies to explore the interactions between crop, climate and management in agriculture systems to provide policy and decision making for food production. A spatial resolution should be determined before using crop model to simulate yield in the regional scale. There is a dilemma in considering the spatial resolution. A high spatial resolution simulation needs more hardware resource and expensive computing cost while a coarse resolution simulation would result in loss of spatial detail of variability. Therefore, exploring the uncertainties of regional simulation of crop model due to different spatial resolutions is essential to application of crop model in large scale. A WOFOST regional simulation platform was constructed at different spatial resolutions to quantify the simulation errors by spatial resolution of model force data. Five spatial resolution meteorological data (5km, 10km, 25km and 50km) were interpolated by the thin plate smoothing spline method which was provided by the software ANUSPLIN. The corresponding resolution soil parameters were extracted from a fine resolution soil database by spatial aggregation. Spatial management and crop cultivar parameters were calculated from observed agro-meteorological sites and then extended to different resolutions by Thiessen polygon method. The simulated yields were compared at different resolutions and statistical yields, and it was found that the average value of anthesis, maturity date, total above ground production and yield at potential yield level and water limited level did not have significant difference. However, there were more extreme values in the high resolution. Each simulation of four resolutions can perform the spatial distribution of crop development. Compared with the regional statistical yields, simulated water limited yield at different resolutions contributed to the variation of statistical yields by 75.4%~85.4%. The correlation analysis between potential yield and water-limited yield and climate factor indicated that the irradiation in the growing stage contributed to 16.6%~29.6% of variability of potential yield, the precipitation in the growing stage contributed to 13.3%~17.8% of variability of water-limited yield. The data storage capacity and computing cost at high resolution, i.e. 5km was a hundred times more than coarse resolution, i.e. 50km. The results provided theoretical and scientific basis for regional application, especially for selecting suitable spatial resolution for regional simulation.