We evaluated the performance of three global evapotranspiration (ET) models using the multiple sets of LAI and meteorological data from 1982 to 2017, and investigated the uncertainty in ET simulations from the model structure and forcing data. The three ET models were the Simple Terrestrial Hydrosphere model (SiTH), Priestly-Taylor Jet Propulsion Laboratory model (PT-JPL ) and MODIS ET algorithm (MOD16). Comparing the observed with simulated monthly ET by the three models over 43 Fluxnet sites, we found that SiTH overestimates ET for forests, but it performed better than the other two models over short vegetation. MOD16 and PT-JPL models performed well for forests, but poorly in dryland biomes. At the catchment scale, all models perform well expect over some tropical and high latitudinal catchments. At the global scale, SiTH highly overestimated ET in tropics, while PT-JPL underestimated ET between 30°N and 60°N and MOD16 underestimated ET between 15°S and 30°S. This study also revealed that the estimated ET by PT-JPL were largely influenced by the uncertainty in meteorological data, while the estimated ET by SiTH and MOD16 were relatively non-sensitive to the forcing data sets. In addition, the results suggested that the long-term variations in estimated ET trend were greatly influenced by the uncertainty in LAI data.