Limited water supply in many parts of the U.S. is a serious problem in agriculture. Policymakers need better tools to devise programs and policies to deal with such shortages. A model is built to estimate irrigation water demand. This model combines a land allocation model with the crop- and region-specific Blaney-Criddle coefficients of net irrigation water requirements. The land allocation model is based on portfolio analysis that combines measures of risks and returns, but it also allows for agronomic and other influences. Two economic methods a simple statistical method and an alternative method using futures prices and a modified weighted average of past yields are employed to generate price and yield expectations. The better of the two methods, based on the root mean square error criterion, is used in the land allocation model to generate measures of risks and returns. The overall model allows for the effect of economic factors on irrigation water demand. This is important because different crops require different amounts of water for their proper growth, and so the crop mix affects water demand in differing degrees. The model is applied to analyze a program used in Georgia to conserve agricultural water use by taking bids from farmers to reduce irrigated acreage. Results from this analysis show that accounting for changes in the crop mix due to expectations about risks and returns for major crops can have significant effects on estimates of water savings from reducing irrigated acreage. This model should allow policymakers to better calibrate acreage reduction programs to meet targeted levels for reductions in agricultural water use.