Abstract: Monitoring the long-term changes of water area is of great significance for the flood and drought monitoring in the East Dongting Lake. In this paper, the remote sensing data set of Landsat series Collection 1, which has been corrected by radiation consistency from 2000 to 2017, is used to extract the long-term sequence water surface area data of the East Dongting Lake area. Four common water extraction methods (single band threshold, multi band spectral relationship, water index, support vector machine) are used under ENVI software to carry out the comparison of the best or most suitable water surface area extraction method in study area. And then, the annual and inter-annual changes of the water surface area of the East Dongting Lake in the past 18 years are studied, also the correlation between the water surface area and the water level of Chenglingji hydrological station is discussed.Our results show that: (1) The remote sensing data set of Landsat collection 1 has a high spatial and temporal resolution, which can meet the requirements of long-time series water surface area extraction in our study area, also the accuracy of water surface area extraction is high regardless of the flood period or water lowering period; (2) During our study period, the average value of multi-year water surface area extracted by support vector machine of the East Dongting Lake is 569.9 km2 , which fluctuates greatly in wet season, but changes little in dry season; (3)There is a close relationship between the accumulation water area change and the water level change in Chenglingji hydrological station. Both linear model and polynomial model can accurately describe the relationship between them; (3) Influenced by the topographical distribution characteristics of the East Dongting Lake, the relationship between the water area and water level is relatively close in the wet season, while the relationship between the two is significantly reduced in the dry season.