Soil forms an important carbon stock in nature but the stocks are dynamic and changes in land-use, land use management and climate can all have significant impacts. The dynamics of the SOC (soil organic carbon) are complex and imperfectly understood and although scientists have developed models for predicting future changes, most have only been validated using individual site data. The National Soil Inventory for England and Wales includes a unique national dataset quantifying changes in topsoil organic carbon under arable, grassland, semi-natural vegetation and woodland (deciduous and coniferous) between the periods 1978-1983 and 1994-2003. Data is available from approximately 2,000 data points located at grid intersects across England and Wales. We used CENTURY 5 (Colorado State University, USA) model to predict changes in SOC between 1978 and 2000 using input weather data for 1978 – 2000 from the UK Meteorological Office and soil property input data derived from the National Soil Inventory (1978-1983). Predicted changes in soil organic carbon from the model simulation was then compared with measured values from the re-sampled Inventory data from 1994-2000. This provided valuable information on the utility of these models with respect to predicting national-scale changes in soil organic carbon stocks, as opposed to predictions for individual sites. Once this utility was established, CENTURY 5 was used to predict national-level climate change-induced changes in soil organic carbon based on the UKCIP02 scenarios for the 2020's (2011 – 2040), 2050's (2041 – 2070) and 2080's (2071 – 2100), each of which comprise four emissions scenarios (high, low, medium-high and medium-low), associated with different socio-economic scenarios. The paper will present the results of model validations based on the National Soil Inventory data, focussing on 10 soil series representing soil types with different water regimes, base status, depth and texture from the 1:250 000 scale National Soil Map database representing a combination of the four different land use categories (arable, permanent grassland, semi-natural grassland and woodlands). CENTURY5 will be used to simulate SOC changes in all four land use classes. Within CENTURY 5, soil organic carbon is fractioned into 4 carbon pools – surface active, fast, intermediate and slow. The soil organic carbon fractionation is different for the various landuse and will be apportioned according to citations in the literature. The bulk organic carbon data from the national datasets has been subdivided into these fractions. Results of the estimated climate change-induced changes in soil horizon organic carbon based on the UKCIP02 scenarios for the 2020's, 2050's and 2080's will also be presented for the same soil series and land use combinations. CENTURY 5 was able to model the changes in SOC in arable, permanent grassland and semi-natural grassland. However when in some instances, CENTURY 5 was unable to perform well for these three landuses, alternative approaches were used to model by altering some parameters (e.g. drainage and anaerobicity factors). The future SOC trends tend to be rather similar to current climatic conditions. Thus for future conditions, in arable and permanent grassland, there seems to be a possible slight increase and constant decline respectively. The SOC trends in semi natural grassland showed a constant increase for future climatic conditions. CENTURY 5 did not perform well for woodlands. The management practices in woodlands are rather localised and site specific which may have resulted in CENTURY 5 unable to model satisfactorily the changes in SOC. Model efficiency was calculated to indicate the reliability of the modelled value compared to the measured data. Root mean square errors and correlation coefficients were also calculated to determine the statistical validity of the modelled results.