Abstract: Taking Santun River irrigation district of Changji Hui Autonomous Prefecture（hereinafter referred to as Changji Prefecture）in Xinjiang Uygur Autonomous Region（referred to as Xinjiang）as research area, hydrological analysis module of ArcGIS was used to extract river network water system of GDEMV3 30 m resolution digital elevation model, so as to obtain river network information such as flow direction, cumulative amount of confluence, hierarchical river network and sub-basin boundary．AI Earth deep learning algorithm was used to classify and extract Sentinel-1 SAR GRD remote sensing images．Different from traditional operation of using ENVI remote sensing image processing software to identify and classify surface objects of images in study area, this study prefered to use AI Earth to classify surface objects of Sentinel-2 L2A image data by using supervised classification algorithm, and then draw a canal system map to clearly and vividly show complex water structure of study area．The 17-grade Tiff data of GDEMV3 30 m resolution digital elevation data, field sampling point data, and selected area of earth map with 0.862 m/pixel and 672 tiles were processed to make canal system topography three-dimensional map of sampling point, and to deeply explore features of drainage geomorphology from a three-dimensional perspective to provide a more comprehensive perspective for the study．Difficulties encountered in course of experiment and solutions were discussed．