The crustal structure under the Tibetan Plateau is quite different from that given by the commonly used Preliminary Reference Earth Model (PREM, Dziewonski and Anderson 1981). We investigate the effects of such differences on inversion results of water trend rates in the area using simulated GRACE (Gravity Recovery and Climate Experiment) and GPS data. When using simulated GRACE data for the inversion, the effects of crustal differences are negligible, confirming the validity of using PREM for the inversion of water trend rates from current and even follow-on GRACE data. However, when using simulated GPS data, effects of crustal differences are very prominent suggesting that an Earth model with a realistic crustal structure instead of PREM should be used for this case.