The RiceGrow model, a process based on the rice growth model, was developed from systematic and comprehensive studies on aspects of the rice cultivation system, such as the mechanisms and relationships among rice growth, climate, soil, and management technology (Fig. 1). The model is composed of seven submodels, including phenology, photosynthesis biomass production, biomass partitioning, yield, quality, and water/nitrogen relationships. Eight genotype-specific parameters relating to rice yield were used: temperature sensitivity, photoperiod sensitivity, optimum temperature, intrinsic earliness (for phonological development), basic filling factor, maximum assimilation rate, potential partitioning index for panicle, and potential relative growth rate for LAI. The submodel equations of the RiceGrow model have been exhaustively described in a previous publication by Tang et al.32 The genetic and cultivar ecotype parameters of RiceGrow were adjusted by the trial and error method33 according to characteristics of the rice cultivars and experimental treatments from the previous experiments,32 which ensure the good predictability and applicability of RiceGrow at various ecosites. The RiceGrow model also could reliably simulate the dynamic processes of rice growth, development, and yield with data from experiments 1, 2, and 3 by validation analysis. The key model state variables LAI and LNA were selected as the coupling parameters for RiceGrow and RS, which not only could express rice growth conditions, but were also directly related to the final grain yield. In addition, data for both of these parameters could be retrieved by RS successfully.34,35 Sowing date, seeding rate, and nitrogen rate, which are generally difficult to obtain accurately at the regional scale, were chosen as the assimilating parameters of the RiceGrow model.