Giant JunCao (Pennisetum giganteum. sp.) is a kind of grass that has been and will continue be widely planted in arid and semi-arid areas of China. The evapotranspiration (ET) dynamic is one of the key concern and unsolved problem, which further constrains the development of topics including irrigation program, water-saving methods and potential, drought resistance performance, and yield prediction. Based on pot control experiment of Giant JunCao, the main process and influence factors of ET were revealed based on SiB2 simulation and regression analysis. Results showed that ET of Giant JunCao, which was dominated by meteorological condition, fluctuated within the range of 0.5 ~ 6.9 mm/day. Among them, ET exhibited unimodal relationship (first enhanced and then reduced) with air temperature, sunshine duration, and vapor pressure deficit, and linearly decreased, increased, and increased with the boost of relative humidity, short-wave radiation, and net radiation, respectively. ET showed different daily change patterns in successive sunny and cloudy days, and mainly affected by short-wave radiation and net radiation. The established multivariate regression relationship between ET of Giant JunCao and meteorological data could explain 99% variance of ET with an error of 3.38±7.64 mm.