Crop models are potentially valuable tools to predict efficient use of water in crop production because they mechanistically consider soil water balance, crop C and N balance, and climatic and crop factors as they change dynamically during the season. The objective of this paper is to evaluate the accuracy of prediction of growth and yield of the CSM/DSSAT crop growth models for several crops such as maize, soybean, peanut, dry bean, and faba bean grown in water-limited environments. The CSM/DSSAT models have several evapotranspiration options including Priestley-Taylor and FAO56 which will be evaluated here. Accurate prediction of water use requires model ability to predict root proliferation and root water uptake at various depths. Also important is ability to accurately predict crop leaf area index over the season to predict dynamic crop coefficients for transpiration. The models implement water stress effects on photosynthesis, leaf expansion, assimilate partitioning, vegetative and reproductive development. In this presentation we describe current abilities of the CSM/DSSAT models by comparison to observed data on soil water balance, evapotranspiration, crop growth and development of several crops over time. Work of Sau et al. showed that the Priestley-Taylor method for predicting climatic potential transpiration is reasonably good, but can be improved on, as it over-predicts ET for cool environments. The FAO-56 equation is the best in our experience, which agrees with findings of most recent FAO manual. Sau et al. found that extinction coefficients for effect of leaf area index on transpiration needed to be reduced in the models, changes that have since been made in the V4.02 release.