Wheat production is expected to be affected by climate change through changing components of the crop water balance such as rainfall, evapotranspiration (ET), runoff and drainage. We used the Agricultural Production Systems Simulator (APSIM)-wheat model to simulate the potential impact of climate change on field water balance, ET and water use efficiency (WUE) under the SRES A2 emissions scenario. We ran APSIM with daily climate data statistically downscaled from 18 Global Circulation Models (GCMs). Twelve soil types of varying plant available water holding capacity (PAWC) at six sites across semi-arid southeastern Australia were considered. Biases in the GCM-simulated climate data were bias-corrected against observations for the 1961–1999 baseline period. However, biases in the APSIM output data relative to APSIM simulations forced with climate observations remained. A secondary bias correction was therefore performed on the APSIM outputs. Bias-corrected APSIM outputs for a future period (2021–2040) were compared with APSIM outputs generated using observations for the baseline period to obtain future changes. The results show that effective rainfall was decreased over all sites due to decreased growing season rainfall. ET was decreased through reduced soil evaporation and crop transpiration. There were no significant changes in runoff at any site. The variation in deep drainage between sites was much greater than for runoff, ranging from less than a few millimetres at the drier sites to over 100 mm at the wetter. However, in general, the averaged drainage over different soil types were not significantly different between the baseline (1961–1999) and future period of 2021–2040 (P > 0.05). For the wetter sites, the variations in the future changes in drainage and runoff between the 18 GCMs were larger than those of the drier sites. At the dry sites, the variation in drainage decreased as PAWC increased. Overall, water use efficiency based on transpiration (WUE_T) and ET (WUE_ET) increased by 1.1 to 1.6 and 0.7 to 1.3 kg ha−1 mm−1, respectively, over the baseline historical climate. Significant relationships between changes in wheat yield and PAWC were only seen at three sites. At the dry sites, the impact of a future climate under a soil of high PAWC was less than that under one of low PAWC. Conversely, the opposite response was seen at two wetter sites, highlighting the importance of PAWC and rainfall in determining the interactive response of crops to primary components of the water balance.