Abstract: Food is the foundation for any production and subsistence. It is important for making national grain policy to estimate yield of winter wheat accurately. In this paper, with the character of BEPS (boreal ecosystem productivity simulator) in simulating vegetation net primary productivity, its applicability and limitation were analyzed in estimating yield of winter wheat. According to the characteristics of winter wheat and forest canopy, a hypothesis about horizontal homogeneous, vertical layered was made. Based on the radiative transfer equation, two-big-leaf model in BEPS model is modified into multilayer-two-big-leaf model. Meanwhile, net primary productivity of winter wheat was converted into yield using harvest index. So the modified BEPS was applied to estimate yield of winter wheat in North China Plain. The simulated yield agrees well with observations from agro-meteorological stations and the R2 reaches to 0.817. This study demonstrates that modified BEPS can be used to estimate yield of winter wheat in North China Plain if remotely sensed imageries are available.