Evapotranspiration (ET) is an important term of surface water-energy balance, and is a significant indicator of crop growth conditions and yield. Compared to the traditional evapotranspiration calculation methods, remote sensing technology is relatively economic, applicable and effective, and it has obvious advantages in monitoring evapotranspiration of heterogenous surface. Many remotely sensed data based models have been developed for regional evapotranspiration estimation during past years. Several frequently-used evapotranspiration models based on remote sensing data and their latest research progress were systematically reviewed first, which were categorized into five types: empirical statistical models, remote sensing models based on conventional methods, surface energy balance models, temperature-vegetation index feature space models, and land surface models and data assimilation. And then, different validation methods including lysimeter at several meters scale, bowen ratio and eddy flux tower at tens to hundreds of meters scale, Large Aperture Scintillometer (LAS) at hundreds to thousands of meters scale, and runoff observation at basin scale were summarized and analyzed. Finally, some existing problems, possible solutions of estimating regional evapotranspiration by remote sensing were analyzed in brief, and the research tendency was prospected. Synergic inversion of surface parameters based on multi-source remote sensing data, evapotranspiration model improvement and multi-model integration, land surface process and data assimilation, the spatial representativeness of observed latent flux and scale problem in the evapotranspiration estimation and validation need to be further studied in the future.