The optical attenuation characteristics of water directly affect the clarity of lake and the survival of submerged macrophytes. Using the way of remote sensing to obtain the optical attenuation distribution characteristics of lakes can greatly improve the efficiency. Lake Daye, located in Huangshi City, Hubei Province, is a typical shallow lake in the middle reaches of the Changjiang (Yangtze) River. Based on the in-situ survey data from 2017 to 2019, the remote sensing inversion model of Lake Daye water optical attenuation coefficient (K_d) is developed by using Landsat 8 OLI images, and the multi-year temporal and spatial distribution characteristics and driving mechanism of K_d in Lake Daye are analyzed, so as to provide reference for the restoration and management of Lake Daye basin. The results show that:the band ratio quadratic model was the best model (K_d=9.61 (B1/B4) ²-2.41 (B1/B4) -6.40), and its accuracy was R²=0.79, MAPE=23.9%, RMSE=0.89 m^-1. The main influencing factors of K_d of Lake Daye were suspended solids and turbidity, followed by chromophoric dissolved organic matter and chlorophyll-a. Moreover, the K_d value of Lake Daye decreased gradually from west to east in space, but it was also affected by wind speed, land-based input, human activities and microbial activities in some parts. The average of K_d is highest in summer and lowest in winter. Since 2013, the K_d value of Lake Daye has shown an overall downward trend.