In this study， variation characteristics of surface fluxes were analyzed by using the eddy covariance observations from four stations of Pailong， Danka， Kabu， and Motuo in the southeastern gorge area of Tibet from November 2018 to October 2019.Pailong Station is located at the entrance of the Canyon in Southeast Tibet， Danka Station is in the middle section， and Kabu Station and Metuo Station are located at the end of the Canyon in Southeast Tibet.Results show that monthly averaged daily latent heat flux is greater than sensible heat flux at night， and it has a single peak during the day.The sensible heat fluxes at Pailong and Danka stations are stronger from November to April and become weaker from May to October.The changing trend of monthly averaged daily sensible heat flux at Kabu station is fluctuating.Sensible heat flux and latent heat flux at Motuo station have the same variation characteristics， with a characteristic of a single peak change.Latent heat fluxes increase first and then decrease at all four stations.Seasonal variations of soil heat flux are obvious， characterizing positive values in spring and summer and negative values in autumn and winter.And the diurnal variation of soil heat flux in Pailong station was the strongest among all stations， and that of Kabu station was the weakest.The diurnal variation intensity of net radiation flux is summer>spring>autumn>winter.The downward and upward longwave radiation fluxes are the largest at Kabu Station and the smallest at Danka Station at the same time they are the largest in summer.The peak of diurnal longwave radiation flux appears later than the shortwave radiation flux.Energy closure rates of Danka， Pailong， Motuo， and Kabu stations are 70.86%， 68.91%， 69.29%， and 67.23%， respectively.In addition， in these four sites the degree of summer closure is high， while the degree of winter closure is low.Latent heat fluxes and soil heat fluxes increase， and sensible heat fluxes decrease as increasing precipitation at the four stations.The sensible heat flux and soil heat flux respond synchronously to precipitation changes， and the changes in latent heat have a significant lag in response to precipitation changes.