The global climate is changing due to constant increase in atmospheric greenhouse gases.The latest climate simulation result indicates that even more drastic change will take place in future global climate,which might exert significant impacts on many production sectors,particularly agriculture.Agriculture is a fundamental production sector for society,as China is an agricultural country with a large population,climate change will bound to exert multiple impacts on Chinese agricultural production.Therefore,using clinmate change projects and crop models to acquire a understanding of the impacts of climate change on Chinese agriculture,especially on winter wheat,is extremely helpful to policy makers and international agencies.CERES-Wheat,a dynamic process crop growth model,has been calibrated and validated for current production at ten sites in the major winter wheat-growing region of Huang-Huai-Hai Plain in China.With the aid of the two Global Climate Models(GISS and Hadley),the present study simulates the production changes under IPCC SRES A2 and B2 climate change scenarios.The simulations consider the impacts on rainfed and irrigated winter wheat with and without CO2 fertilization.The results indicated the possibility of significant impacts of climate change on winter wheat production in this region,with marked differences between rainfed and irrigated production.The conclusion is that an overall reduction of the rainfed wheat yield was observed in the study area,the yield reduction extent is greater in the western part and smaller in the eastern,taking no account of direct CO2 fertilization effect while the irrigated wheat yield keeps the current level with the increase in irrigation water amount but without considering the CO2 fertilization effect.Meantime both the rainfed and irrigated wheat yield increased largely with the consideration of the CO2 fertilization effect,but the increasing extent of rainfed wheat is greater than the irrigated wheat because the CO2 fertilization effect improves significantly water use efficiency and reduces the irrigation water amount.Our simulation results also showed that the selection of GCM could significantly affect final predictions,especially in relation to precipitation signals.By contrast,we found that the selection of emission scenario(A2 vs.B2) was less important in determining final impacts.