Abstract: It is difficult to study water saving irrigation in saline soil due to more influencing factors. However, the root-water-uptake rate is important for water saving irrigation in saline soil. Therefore, the root-water-uptake rate in saline soil was studied. According to one dimensional vertical water flow equation with a volumetric sink term (represent the water uptake by plant roots), the finite difference method was used to calculate the root-water-uptake rate with field data under water saving irrigation. The accuracy of the root-water-uptake rate was tested. The Artificial Neural Network was introduced to develop the soil-root water extraction model, and the topology structure was 15∶12∶5. Results show that the main water uptake by the sunflower roots located the 0～50 cm layer, and the peak rate of water uptake moves between 20 and 40 cm soil layer. The calculating precision of the root-water-uptake is higher. The BP neural network model for the root-water-uptake rate by sunflower in saline soil has a reasonable precision.