Crop water production functions require precise knowledge of crop productivity across a wide range of soil water availability (WA). Field studies have demonstrated the yield response of maize (Zea mays L.) to WA, but the relationship has not been quantified across various precipitation years in the North China Plain (NCP). This study was conducted to investigate the effects of the simulated interannual variability in precipitation on maize yield in a winter wheat (Triticum aestivum L.) --summer maize double-cropping system in the NCP, in 2014 and 2015. A rainfall simulator system was used to simulate contrasting precipitation years, including wet, normal, normal-dry, and dry years. A fixed amount of 70 mm irrigation was applied at different growth stages of maize, including at planting (I70), at planting and tasseling (VT) stages (I140), and at planting, VT, and milk stages (I210). The objective was to quantify the relationships among crop production, water use, and WA for maize. We found that I210 improved soil water storage (SWS, mm) by 19 to 36% in the midseason of normal-dry and dry years but had no effect in normal and wet years, compared with I70. Crop evapotranspiration (ETc, mm d-1) was least in the dry year with the I70 treatment and differed with maize growth phases, with the peak occurring midseason for maize in both years. Regression analyses showed that ETc was negatively related to SWS but positively related to WA. A quadratic function described grain yield responses to WA and ETc, with total WA of 478 mm and daily mean ETc of 5.1 mm d-1 producing greatest grain yield. With yield response function acquired, we conclude that target maximum yield of 11.6 Mg ha-1 can be achieved after probability of exceedance of precipitation equals 18 with irrigation water requirement applied. [ABSTRACT FROM AUTHOR]