Abstract: In this paper,the distribution and total amount of evapotranspiration (ET) is estimated under the diverse landscape over the Heihe river basin in Northwestern China in July 2004 by using MODIS data and meteorological data based on the aerodynamic method and the time upscaling scheme.Taking advantage of the high resolution Landsat-TM data,according to the contrastive analysis,we find that the influence of the heterogeneous underlying earth's surface on the estimation of water and heat exchange on the land surface by using the high resolution remote sensing data is obviously smaller than that by using the middle resolution ones.By comparing the estimates of ET with the observation data collected from "JTEX" experiment we conclude that,by using the high resolution data,the parameterization algorithm used in this paper can derive daily ET with a relative error of 10%.After analyzing the distribution of the derived monthly ET over the basin in July 2004,it is concluded that the mountainous area above 2000 m exhibits the ET twice more than that the area fronting mountains below 2000 m elevation; and the oasis and crop area take on an ET 12 times more than that of the desert and Gobi region over the area fronting mountains,and the former is the most significant region that consumes land surface water over the mid-downstream area in summer.