The Qiandongnan Miao and Dong Autonomous Prefecture is located in the southeast of Guizhou Province. It is a crucial forest area in China and a significant ecological barrier area in the upper reaches of the Yangtze River and the Pearl River. Based on land use/land cover (LULC) data and socioeconomic and field survey data from 1990 to 2015, the study analyzes the spatiotemporal changes of ecosystem health (ESH) in Qiandongnan Prefecture by using the "Pressure-State-Response" model, "Vigor-Organization-Resilience-Services" theory frame and spatial analysis model. In this study, we selected urbanization and terrain index to detect the impacts of different factors on its ecosystem health. In terms of time distribution, ecosystem health index of Qiandangnan prefecture changed from a "sub-healthy" level to a "threshold" level from 1990 to 2015, with a slight increasing from 1990 to 2005 before plunging at 2005. The highest area in ecosystem health level occupied by "sub-healthy" (42%-45%), with the level of "critically healthy" accounting for 30%-35%, "healthy" 15%-22%, "unhealthy" 0.1%-0.4% and "sick"<0.1% in decrease order. In terms of spatial distribution, Taijiang County and Jianhe County, located in the center area, as well as Sansui County and Jinping County, located in the northeast, were at relatively health levels. The counties located in southern, eastern and northern regions were mostly in a "threshold" healthy state, such as Rongjiang County, Liping County, Zhenyuan County, and Cengong County. However, the lowest healthy area belonged to the northwest area like Kaili City, and its health index decreased from 1990 to 2015. Studies on spatial correlation showed that there was a significant positive spatial correlation between the terrain index and ecosystem health during the period of 1990-2015, and the areas, featured by "low-low" agglomeration and "high-high" agglomeration, presented a fluctuating trend of "decrease-increase-decrease". While there was a significant negative spatial correlation between urbanization and ecosystem health only in Kaili, capital of study area. And the results indicated that urbanization factors such as population density, per capita GDP, and the proportion of construction land were the main reasons affecting the ecosystem health in places with high levels of urbanization. However, the ecosystem health was greatly affected by natural factors in areas with a low urbanization process, for example, the terrain gradients. The research provides a theoretical basis for the management and the sustainable development strategy of the ecosystem of Qiandongnan Prefecture.