Abstract: Drought has posed an ever-increasing impact on agricultural production in recent years. A crop model has widely been an effective tool to explore the effects of drought on agriculture. Accurately simulating water stress in a crop model is a key step to assess the effect of drought impact on crop growth and development in field. In this study, three algorithms of water stress were integrated into a standard platform, where three kinds of water stress models were composed of: average Soil Water Content (SWC), Water Supply to Demand ratio model (WS/WD), and Actual to Potential Transpiration ratio model (AP/TP). The parameters of models were calibrated and verified using field observation data in the irrigation experiment from 2017 to 2019 in Wuqiao, Hebei Province, and literature data of irrigation experiment from 2008 to 2009 and 2013 to 2016. Five irrigation scenarios were designed, including rainfed, one irrigation (75 mm), two irrigations (150 mm), three irrigations (225 mm), and four irrigations (300 mm). The results showed that simulated values fully represented the measured ones with a reasonable error range under different water stress models. Therefore, the normalized root mean squared error (NRMSE) values of root depth, root biomass, anthesis, and maturity were 10.2%, 17.1%, 1.2%, and 1.0%, respectively. The NRMSE values of leaf area index, above-ground biomass, soil water content, and yield under three models ranged from 26.6%-33.1%, 14.0%-16.5%, 5.1%-8.8%, and 5.4%-7.7%, respectively. There was also a difference in the occurrence time and severity of water deficit that was simulated by three water stress models during the wheat growing season. Nevertheless, there was a consistent trend of interannual water stress factors. Furthermore, the factors of water stress simulated by the SWC and the WS/WD model was relatively higher than those by the AP/TP model. The water stress simulated by the three models in wet years was lighter than that in dry years. There was an earlier occurrence of drought that was predicted by the AP/TP model, whereas, the latter by the WS/WD and SWC model. Precipitation during the growing season dominated the variations of water stress factors under rain-fed conditions, which were 56%, 56%, and 39% in the SWC, WS/WD and AT/PT models, respectively. In addition, there were different effects of three water stress models on the winter wheat yield, water use efficiency, and irrigation water use efficiency. Specifically, the grain yield improved greatly, while the water use efficiency increased first and then decreased, whereas, the irrigation water use efficiency decreased under three models, as the irrigation times increased. There was obviously distinguished from the yield, water use efficiency, and irrigation water use efficiency in the three models. Particularly, the trends of irrigation water use efficiency were different under various water treatments. The SWC, WS/WD, and AT/PT models simulated that the yields in four irrigation treatments were 163%, 132% and 92% higher than those of rain-fed treatment, respectively, and the irrigated water use efficiencies under four irrigation treatments were 26.8%, 12.3%, and 40.0% lower than those under one irrigation. The highest water use efficiency simulated by WS/WD was found in the three irrigation treatments in dry years, three irrigation treatments in normal years, and two irrigation treatments in wet years. Correspondingly, the WS/WD model performed the best, while the SWC model was the second, and the AP/TP model was the third, particularly in the water decision-making of winter wheat in Wuqiao County. Consequently, it is vital to fully consider the differences in three algorithms on the wheat growth and development under different water stress, thereby to improving the reliability of crop models on drought impact assessment and water management of winter wheat.