It is vitally important to accurately estimate soil organic carbon (SOC) dynamics modeling uncertainty for successful decision-making in cropland management. In this study, a Beyesian inference using the Markov Chain Monte Carlo (MCMC) method was used to calibrate the parameters and estimate the output uncertainty interval for the Denitrification-Decomposition (DNDC) model in modeling SOC dynamics of the long-term monitored croplands with 22-year rice-wheat rotation history in Yixin County, Jiangsu Province of China. Results indicate that though there is some uncertainty, the DNDC model is suitable for modeling SOC dynamics of the long-term monitored croplands. In cases that input parameters are not certain in quality, Beyesian inference using the MCMC method could be an effective way for automatically calibrating input parameters and estimating uncertainty interval of the modeled SOC. The Beyesian inference using the MCMC method applied in this paper could also instruct the estimation of the SOC dynamics modeling uncertainty at region or country scale.