Delineating the root-water-uptake (RWU) under the salinity stress condition accurately is important for making rational saline water irrigation schedules. However, RWU is very complex and still incompletely understood, especially under the salinity stress condition. In this study, the linear relationship between RWU and root nitrogen mass (RNM) of winter wheat, which was obtained under the optimal water condition (without any water or salinity stress) from the greenhouse experiments, was further verified and applied to optimize the parameters of the salinity stress reduction function (β), and simulate soil water dynamics under the salinity stress condition and in the field. Three saline water irrigation experiments, with winter wheat cultured in nutrient solution (Exp. 1), soil columns (Exp. 2) and field (Exp. 3) were conducted. The results of Exp. 1 showed that the daily transpiration of winter wheat decreased with increasing salinity, but was proportional to the RNM under the same salinity stress. The RWU rate distributions in Exps. 2 and 3 were estimated using an inverse method. The estimated RWU rate was also verified to be proportional to the RNM density under the optimal water condition. The verified linear relationship was then used to optimize the parameters of β, simulate the distributions of RWU rate and soil water content in the soil columns (Exp. 2) and field (Exp. 3). The simulated results matched the estimated RWU rates and the measured soil water contents well. In Exp. 2, the maximal absolute error (MAE), root mean squared error (RMSE) and absolute coefficient of residual mass (CRM) were not more than 0.0021, 0.0062cm 3 cm −3 d −1 and 9.93% between the simulated and estimated RWU rates, and 0.007, 0.004cm 3 cm −3 and 2.12% between the simulated and measured soil water contents, respectively. The relative errors between the calculated transpiration rates from the simulated RWU rate distributions and the measured values were less than 10%. In Exp. 3, the maximum MAE, RMSE and absolute CRM were 0.0012, 0.0007cm 3 cm −3 d −1 and 5.81% for RWU rate, and 0.011, 0.010cm 3 cm −3 and 3.51% for soil water content, respectively. The linear relationship between RWU and RNM of winter wheat should be reliable and rational. The optimization method should be applicable in optimizing the parameters of the salinity stress reduction function, establishing the RWU model and simulating soil water flow under the salinity stress condition in the soil–wheat system.