In order to remove the influence of shadow noise on water extraction from remote sensing images effectively, and calculate the water surface rate in a large area quickly and accurately, this paper proposes a method which uses the UWI(Urban Water Index) index and FROM-GLC10(Finer Resolution Observation and Monitoring of Global Land Cover 10) data based on GF6 image to achieve the refinement of water body extraction results through intersection calculation and small speckle deletion, eliminate the interference of shadow noise on water body extraction effectively, and achieve the rapid calculation of water surface ratio. Zhongshan City is selected as the test area in this paper which is large and the ground features are complex. By comparing the algorithm in this paper, NDWI(Normalized Water Index) algorithm and TSUWI(Two-Sted Urban Water Index) algorithm, the qualitative and quantitative analysis shows that the water body extracted by the algorithm in this paper is continuous and complete, with fewer omissions and errors, which can remove noise interference such as shadows effectively, buildings and roads, and the bestter results have been achieved in the overall accuracy, Kappa coefficient, wrong score error, missing error and other indicators. It is concluded that the combination of UWI index and FROM-GLC10 data is feasible to calculate water surface ratio, which provides a new idea for engineering and large-area calculation of water surface ratio quickly.