Soil texture has been considered a key factor controlling soil carbon (C) accumulation and a positive relationship is generally found between clay content and soil C. However, the effect of soil texture may vary with different management practices. CENTURY model is one of soil organic matter models that have been used widely to study integrated effects of various components of agricultural systems on C sequestration. In this study we applied the CENTURY model to predict changes in soil organic matter occurring in four treatments, such as conventionally till and no till treatments with high chemical inputs, and conventionally till treatments with low and zero chemical inputs and winter legume cover crop at the Long Term Ecological Research site at Kellogg Biological Station, Michigan. We, first, analyzed the sensitivity of the model to soil texture variations for these treatments. Second, we evaluated magnitude of spatial variability in soil C model predictions as a function of spatial variability of soil texture within studied 1-ha plots. The simulation results represented well the changing trend of soil organic carbon for the studied treatments. The model was shown to be sensitive to (clay+silt) fraction in predicting soil organic carbon and the effect of soil texture became more significant with time. Among the studied treatments, no till was the management practice where soil C predictions were most sensitive to texture.