At 21:19 p.m., August 8, 2017, a 7.0-magnitude earthquake struck Jiuzhaigou, Sichuan Province. In a couple of hours, a great deal of the earthquake-related information spread on internet. The earthquake drew much attention from the social media and soon became a hot topic. In this paper we searched those Sina Weibo users who are within the range of 200km from the epicenter, and copied their Weibo data released 24 hours before and after the earthquake event. After cleaning, mining, and classifying these data, we analyzed their characteristics such as quantity, word-frequency, and classification, spatial and temporal distribution. We found that extracting data from the social media would help governments learn overall the post-earthquake information, and on this basis make decisions and arrangements for earthquake relief.