Abstract: Based on the relationship between crops and circumstances,by using system method and technique of mathematical modeling, a dynamic knowledge model system for maize management, which can be used in different places and conditions, was developed. With the incorporation of the soft component characteristics,a component and digital knowledge model-based decision support system for maize management was established on the platform of Visual C++. This system realizedsix major functions——target yield decision, design of pre-sowing technical plan, prediction of regulation indices, real time management control, expert knowledge reference and system administration. Case studies on the target yield knowledge model with data sets of different eco-sites, yield levels of last threeyears, fertilizer and water management levels and so on were conducted, and theresults indicate that this model has wide applicability and good independance, which provides a scientific knowledge system and broad decision-making tool formaize management.