Abstract: Using Sina Weibo as the data source, the ROST artificial intelligence natural language processing technology was applied to analyze the time characteristics, word frequency characteristics, social network semantics, and emotions of the relevant Weibo data for the 4.7-magnitude earthquake in Feidong County, Hefei City, Anhui Province on September 18, 2024. The visualized results presented the comprehensive perception of the public and earthquake public opinion from different perspectives, providing some reference for future earthquake public opinion disposal work. The research results indicate that: ① The Weibo post volume within 48 hours after the earthquake showed a "double M" trend, with the peak of Weibo post volume occurring within 1 hour after the earthquake, and highly followed within 4 hours after the earthquake. In the obtained blog posts, the term "earthquake" appears the most frequently, with words such as "Hefei", "Anhui", "Feidong County", "Zhengan", "Aftershock" appearing thousands of times, and descriptive words such as "crack", "feel", "shake" appearing hundreds of times; ② The overall performance of public emotions is relatively stable and the situation is good, with positive emotions accounting for 64.07%, neutral emotions accounting for 19.47%, and negative emotions accounting for 16.46%.