Abstract: Agricultural irrigation accounts for>65% of water use in the Beijing-Tianjin-Hebei Province Plain. Improving the agricultural water use efficiency will help relieve the pressure on the water resources found in the Beijing-Tianjin-Hebei Plain and promote sustainable development. Based on water use efficiency (WUE), net primary productivity (NPP), and actual evapotranspiration (ETa) from 1980 to 2013 simulated by the VIP model, combined with the annual mean air temperature (Tmean), annual precipitation (Pre), and annual sunshine duration (Sun), the Copula method was used to create five groups of joint probability distributions:WUE and NPP, ETa, Tmean, Pre, and Sun. Conditional probability was calculated based on the hypothesis that WUE was greater than any particular value under low, medium, and high value ranges of each driving factor. The findings showed that the greater the values of NPP, ETa, and Sun, the more likely was WUE to be greater than any particular value. However, the lower the values of Tmean and Pre, the more likely was WUE to be greater than any particular value. The sensitivity of WUE to variation in the value of each driving factor was reflected by the difference of the conditional probability of WUE under high and low value ranges, suggesting that WUE was most sensitive to the variation in the value of NPP followed by those of Sun, ETa, Pre, and Tmean. Comparison of the conditional probabilities of WUE under the same value conditions of NPP, ETa, Tmean, Pre, and Sun showed that a lower NPP clearly suppressed WUE and that improvement in NPP guaranteed a higher value of WUE. Crop WUE is controlled simultaneously by photosynthesis and transpiration, which makes it difficult to ascertain the driving mechanism underlying WUE. Based on the joint probability distribution determined using the Copula method and conditional probability analysis, we concluded that improving NPP when water consumption is controlled may be a more effective strategy than reducing water consumption when grain yield is controlled to adopt in the Beijing-Tianjin-Hebei Province Plain.