It is important to estimate reference crop evapotranspiration precisely in the process of regional water resources and land use balance evaluation. Some methods for estimating reference crop evapotranspiration were validated, which include FAO-24 Blaney-Criddle method, FAO-24 Radiation method, FAO PPP-17 Penman method and FAO Penman-Monteith(98) method. Analysis of the correlation between the calculated reference crop ET with the evaporation measured by 20 cm pan, FAO Penman-Monteith(98) performed the best, with R² value equaling 0.89, whereas values of R² for FAO PPP-17 Penman, FAO-24 Blaney-Criddle and FAO-24 Radiation was 0.88, 0.87, 0.84 respectively. Sensitivity analysis of these models showed that FAO Penman-Monteith(98) model and FAO PPP-17 Penman method were better in performance, while FAO-24 Blaney-Criddle method and FAO-24 Radiation method were worse. Further more, FAO Penman-Monteith(98) model was applied to calculate reference crop evapotranspiration in Gansu Province from 1981 to 2000 year. The results showed that the highest E_TP of multi-year monthly average was in July, and the variation of annual reference crop E_TP was consistent with the evapotranspiration variation of summer season, further more ET was the highest in southeast and the lowest in northwest.