Abstract: The accuracy and convergence of the inverse method to estimate the average root-water-uptake rate, which was proposed by Zuo, et al.(2001), are mainly dependent on the following two factors: the time interval (T) between the two successive measurements of soil water content distributions, and the spatial interval (SI) of the measured data along the soil profile. Examples of the numerical experiment were taken to examine the effect of different T and SI on the accuracy and stability of the method in this paper. In practice, it is recommended to choose 5 d≤T<15 d and SI<20 cm so as to keep the maximal absolute error and the overall relative error between the estimated and the real average root-water-uptake rate within 0.0015 d^-1 and 15%, respectively. Furthermore, the inverse method was applied to simulate the average root-water-uptake rate distributions of winter wheat at seedling stage under water stress. Winter wheat was planted in the soil columns and the experiment was conducted in a greenhouse with 2 water stress levels. The simulated results showed that water stress reduced root-water-uptake of winter wheat remarkably and probably inhibited the roots growth as well. The scope of winter wheat root-water-uptake was adjusted by itself to the lower soil layer where soil moisture was appropriate when soil water supply in the upper soil layer was deficient.