Abstract: Based on integrating the effects of yield potential of photosynthesis and temperature, average yield level of last three years, soil fertility, fertilization and water management level and production technology level on yield increment, the dynamic yield increment index was quantified and a wheat knowledge model for design of target yield under different temporal and spatial environments was developed through knowledge engineering and system analysis method. By quantitatively calculating suitability of variety characteristics to environmental conditions and production requirements, a knowledge model for variety selection was established. Case studies on the target yield design with the data sets of five different eco-sites, three climatic years and average yield levels of last three years, two soil fertility levels and three fertilization and water management levels, and on suitable variety selection models with the data sets of five different eco-sites, normal climatic year and fifteen typical varieties indicated a good performance of the model system in decision-making and wide applicability.