In 2011, Texas experienced the most severe drought in its 100 years history, resulting in the death of more than 300 million trees. On a large spatial and temporal scale (an area of 9×10⁶ ha, a time span of nearly 20 years), this study is based on 1800 forest plots, with about 209,663 trees from four periodic surveys. This study utilized Principal component Analysis (PCA) and Generalized Linear Models (GLM) to characterize the spatio-temporal difference of tree mortality and drought severity and length effects. The causes of the differential tree mortality at cycle and annual level before and after the 2011 drought were analyzed using tree density, basal area, stand age, and survey interval as the competition indexes among trees. The relative contribution of tree death to mortality was analyzed according to four classification criteria, geographic region, tree species group, tree sizes (height and diameter at breast height-DBH) and forest stand origin. The results showed that the Pinus trees suffered the least, with a mortality rate of 7.92%. The mortality rates of trees with low height and small DBH were 29.79% and 26.00%, respectively. Natural forest trees witnessed a higher mortality rate (13.47%) than plantation forests (10.26%). The mortality rate of trees in the Western Gulf Coastal Plains ecological region and southwest region reached the highest after drought with a mortality rate of 22.27% and 13.78%. The effects of elevation and latitude on tree mortality were not apparent. The causes of overall tree mortality patterns in eastern Texas are complex, and differential mortality rates were observed across geographical regions, stand origin, tree sizes and different tree species. Among the four regions, the natural forest had a higher death contribution rate. Small trees contributed the most to the death in both natural and plantation forests. The effect of drought on the natural forest was more significant than that on plantation forests. The death contribution rates in natural forests and plantation forests are different. In the natural forests, the death contribution rates of Pinus and Liquidambar species were lower than those in the plantations. Regression analysis of forest density and tree mortality did not show a significant statistical relationship between them, suggesting tree mortality was triggered by density-independent factors. The results of PCA showed that the effect of drought on tree mortality was greater than that of tree competition and other natural disturbances, and the effect of drought severity and drought length on tree death was similar.