Abstract: Global ecological environment of river is causing increasing concern in government and research communities．Shape of river cross-section is an important parameter to evaluate ecological environment of river．Traditional measurement methods are mostly based at total stations and level stations．This is not only time-consuming and labor-intensive, but also poses high risks in field work．It is therefore urgent to develop a framework to efficiently deduce the shape of the river cross-section．For this study, the Sanjiangyuan region with less human impact was selected．Multi-source remote sensing data coupled with least square and fuzzy clustering methods were used to explore correlation between shape of cross-section and key impact factors．Among all the influencing factors, drainage area was found to be the most important factor driving shape changes in the river section．As drainage area controlled by the section increased, upstream “narrow-deep” channel form gradually changed to downstream “wide-shallow” form．Geographical environment with high soil pores, large silt content, high slope of section location, and low vegetation coverage was found to be more conducive to the formation of parabolic river sections．The relationship functions of section shape fitting parameters (a), control basin area (S) and width-depth ratio (R_WD) were all exponential functions (a=c_1\rme^d_1S, R_\rmWD=c_2\rme^d_2S)．This relationship is suitable for natural rivers in mountainous areas that are less affected by human activities, accurately, to efficiently, and safely obtain the shape of river cross-sections, to provide methodological and data basis for the assessment of water quality, calculation of e-flows, restoration of aquatic ecosystem health as well as conservation of water and soil, being helpful to achieve ecological civilization in China．