For coastal waters, the radiation signal received by a sensor is affected by the high reflectance of the terrestrial land, which can result in a reduced image contrast and significant adjacency effect. Eliminating atmospheric attenuation and the adjacency effect and obtaining accurate water surface reflectance are important prerequisites for quantitative water color remote sensing. Based on the in situ spectral data of the coastal waters of Taihu Lake on April 29, 2016 and a synchronous Gaofen-1 wide field of view (GF-1/WFV) image, the 6S model was used to eliminate atmospheric attenuation, and a nuclear function is used to express the atmospheric point spread function to correct the adjacency effect. The experimental results show that the image after correcting the adjacency effect is clearer with an increased contrast and more abundant information regarding the water body. Compared with the correction results of the 6S model, the average relative errors of the three in situ synchronous samples after correcting the adjacency effect are reduced by 10.8%, 5.24%, and 10.39%, respectively. Correcting the adjacency effect improves the accuracy of radiation detection of the remote sensing reflectance above water surface.