In soil hydraulic models, the input of hydraulic parameters is an important reason for differences in model results, and these soil hydraulic characteristics can be estimatedusing pedotransfer functions (PTF). To analyze the performance of different PTF, the soil water content at seven FLUXNET sites was simulated and predicted using the VGM model and the BC model within the HYDRUS- 1D framework. The results showed that using different PTF led to significant differences in the predicted results. For the VGM model, Rosetta3 was the most stable PTF; for the BC model, Cosby2 was the most stable PTF. Based on the simulation results of this study, different PTF should be compared within a model to avoid the uncertainty caused by PTF selection.