Precise description of crop physiological growth process and accurate estimation of crop yield play a crucial role in social food safety and agricultural production. As a numerical simulation method for monitoring crop growth,the crop growth model has proven a powerful tool for crop growth analysis, yield simulation and scenario prediction. However,under the current background of climate change, the uncertainty of climate trend and the increase of extreme weather events have affected grain production largely. Therefore,how to simulate crop growth accurately under the climate change scenario has become a hot topic currently. Except the uncertainty of climate change,crop models are a simplified form of actual crop growth process. Thus these models have an imperfect physical structure. Additionally, randomness in data collection and regional heterogeneity of initial conditions can lead to poorer uncertainty in simulation results. Hence,based on a review of the state of crop models in China and abroad,the uncertainty of crop models and the impact of uncertainty of climate change on the models were discussed. Then,the types and characteristics of model-data fusion methods were outlined. The methods used to reduce the uncertainty of crop models in a regional scale were summarized.