Located in the subtropical humid monsoon climate zone, Lake Dongting is significantly experiencing hydrological regime changes at temporal and spatial scales. In this paper, the data of water level at six stations (Chenglingji, Lujiao, Yingtian, Yangliutan, Nanzui, Xiaohezui) in Lake Dongting and streamflow of Sankou, Sishui, Chenglingji, Yichang, Luoshan stations were collected to study the temporal and spatial variations of water level in Lake Dongting. Both water level and streamflow data are daily observation from 1985-2014. Water level of the lake was calculated by Thiessen Polygon. Wavelet analysis, cluster analysis and geostatistics were used to reveal temporal and spatial variations of water level in Lake Dongting. Results indicate that the water level shows typical seasonal features, and its annual fluctuation has multiscale of 28 a and 22 a in Lake Dongting. There are two clusters of the spatial distribution pattern in Lake Dongting, one is Xiaohezui, Nanzui, Yangliutan (Group 1), and another is Chenglingji, Lujiao, Yingtian (Group 2). The magnitude of spatial autocorrelation in different periods is wet season > retreating season > rising season > dry season. The driving factors of temporal and spatial variation of water level in Lake Dongting were revealed by multiple stepwise regression model among two kinds of water level and runoff in four seasons. The hydrological regime alteration in Yangtze River is the main factor for Group 1. In different seasons, the driving factors of Group 2 are different,including the discharge of lake and the hydrological regime of Yangtze River. The difference is mainly caused by the relationship between Lake Dongting and Yangtze River as well as the flooded and exposed bottomlands in different seasons. The study is essential for protecting the ecosystem of Lake Dongting and reasonably regulation, management and utilization the water resources.