Greenhouse irrigation optimization decision support system (GDSS) is developed for the irrigation management of greenhouse crops in the northwest arid area of China. GDSS forecasts on relative yield and aids to develop irrigation schedule in terms of growth periods, comprehensively considering soil, crops and water supply conditions. The system consists of three modules. The database module stores all kinds of data using the Access database. The model module includes optimization models of greenhouse crops under insufficient irrigation based on uncertainty. These models are programmed by lingo, which can be invoked through internal interface. The man-machine dialogue module is designed with the principal of user control, user-friendly, visuality, usability, conciseness and uniformity. The GDSS can provide the decision makers the alternative decision making under uncertainty.