Abstract: Abstract: It is generally accepted that the model parameters calibration is an essential step before application. However, the traditional methods of obtaining a set of "optional parameters" based on a certain number of observations, fail to represent the uncertainties and get reliable estimates. To overcome this problem, we introduced the DE-MC (differential evolution Markov chain) algorithm with snooker update into the WOFOST (world food studies) model parameters calibration. The main objectives were: 1) to calibrate the WOFOST model by DE-MC algorithm with snooker update; 2) to evaluate the uncertainty of the model parameters after calibration; 3) to evaluate the uncertainty of the model outputs after calibration. Observational data including LAI (leaf area index) in different growth stages and the final yield of winter wheat in Zhengzhou Agrometeorological Experimental Station, were used to calibrate WOFOST model at potential mode with this algorithm. The crop parameters related to accumulated temperature were calculated directly from the crop phenological development date and near-surface temperature. The remaining crop parameters were analyzed by Sobol method. Taking Sobol global sensitivity index greater than 0.05 as the threshold for sensitive parameters, 14 parameters were then selected. The calibrated parameters were defined as the uniform distribution over their value interval. The likelihood function represented the mismatch of the model output with the measured observations, by which the parameters' priori distribution converted to posterior distribution. The likelihood function for yield was set as a Gaussian distribution with the observational data as expected value and 10% of the observational data as standard deviation. Similarly, the likelihood function for LAI was set as a multidimensional Gaussian distribution with observational LAI as the expected vector and a diagonal matrix as the covariance matrix. At last, we found that: 1) Compared with the simulation results with model default parameters, the LAI simulation accuracy could be increased by 51.40%-53.07% after the parameter calibration, and the yield simulation accuracy is improved by 8.25%-8.88%; 2) The posterior distribution of life span of leaves growing at 35 Celsius (SPAN), specific leaf area at development stage of 0.7 (SLATB070), specific leaf area at development stage of 0.4 (SLATB040), maximum CO2 assimilation rate at development stage of 1.3 (AMAXTB130), and specific leaf area at development stage of 0 (SLATB00) could be approximated as a Gaussian distribution with SPAN having the minimum uncertainty; 3) Running model with the posterior parameters set, the uncertainty of simulated LAI from the three-leaf stage to the re-greening stage and from the jointing stage to the heading stage is larger; the uncertainty of the simulated yield increases with time before the milky ripe stage, stayed unchanged until maturity. It was concluded that DE-MC with snooker update was an effective algorithm for parameters estimation of WOFOST model, the calibration had the potential to reduce the uncertainty of the model parameters and the calibrated model was able to model the observational data with some degree of skill. The calibration data used in this study was the observation data from the agrometeorological site. Although the crop varieties were consistent, the inter-annual cultivation management measures were not strictly controlled (eg.: seeding density), and the accuracy of the calibration results might be affected. Subsequent experiments can be carried out in more areas using experimental data with quantitative control to further verify the effectiveness of this algorithm.