Increasing concentrations of greenhouse gases are expected to result in global warming which will affect crop production. Crop modelling is a useful tool for assessing the impact of climate change on crop production. The objective of this study was to select an appropriate model for climate change studies. Five simulation models, EPIC, CERES, Century, Sinclair and Stewart, were assessed using data from a long-term experiment begun in 1911 on a clay loam (Dark Brown Chernozem) soil at Lethbridge, AB. Yields predicted by the five models were compared with actual spring wheat yields in continuous wheat, fallow-wheat and fallow-wheat-wheat rotations. The EPIC model gave the best simulation results over all rotations and the most accurate predictions of mean yields during droughts. It was concluded that the EPIC model had the greatest potential for assessing the impact of climate change on wheat yield. The Stewart model was the most accurate for unfertilized continuous wheat and fallow-wheat. The Sinclair model was most accurate for fertilized fallow-wheat and CERES was the most accurate model for fertilized continuous wheat. The Century model simulated average yield accurately but did not account for year-to-year variability. Key words: Global warming, crop simulation, spring wheat yields