Through analyzing the newest research results about nitrogen management of cotton and experimental results, and on the basis of the different effects of cotton yield, plant height and leaf area, a dynamic knowledge model for decision making on total nitrogen and their distribution among main growth stages of cotton under different environments and yield targets was developed with the principles of nitrogen balance and by integrating the factors of climate, soil, yield and so on. A comprehensive and intelligent computer system for cotton management was established using VC++6.0 and SQL Server 2003.