Soil carbon is an important component of global carbon cycle. And soil organic carbon ( SOC ) storage is so huge that slight change of it can contribute to significant change of global CO_2 concentration. Accurate evaluation of SOC dynamics temporally and spatially is of great significance to the decision making for the global climate change. Process-based models are recommended means to estimate SOC dynamics regionally. However, input parameters in each modeling unit which were upscaling from data with low temporal and spatial resolution will always introduce uncertainty to the model outputs. Thus, identifying the input parameters to which the modeled SOC is sensitive can guide the future survey design to reduce the uncertainty of it. Recently, many researches about sensitivity analysis have been done. However, little research has been done about sensitivity of CENTURY model's output SOC to model input parameters in China, especially those parameters about agriculture management which are differed over the whole country. In our study, CENTURY model was validated in a long-term monitoring site located in Licheng County, Shandong Province of China. The performance of CENTURY model proved that it worked well in typical cinnamon soil in Shandong province represented by this long-term monitoring site. Global sensitivity analysis was used to quantify the sensitivity of modeled SOC to model input parameters in Shandong Province of China. The simulation of typical cinnamon soil in Shandong province indicated that SOC density under this circumstance increased from 1. 9kg C/m~2 in year 1987 to 2. 9kg C/m~2 in 2007. Sensitivity analysis results indicated that during the simulation period, the sensitivity of the modeled SOC varied over time and differed according to the input parameters. Six out of twelve input parameters are of influence to the model output SOC,which are percentage of harvest straw, manure addition, temperature addends, soil clay content, precipitation multiplier and potential aboveground monthly production coefficient for maize in order of the total effect. Percentage of harvest straw and manure addition are major uncertainty sources of modeled SOC with the highest total effect (the total effect of them are 0. 37 and 0. 20,respectively). Therefore, in order to reduce uncertainty in the modeled SOC, the amount of straw residue and manure addition should be emphasized in the agricultural management survey. Reasonable measures to reduce uncertainty in CENTURY's output SOC can be conclude according to the findings of this study. And this study also useful for the uncertainty assessment of modeled agricultural SOC's temporal and spatial dynamics at national scale.