In　December　2022,　China　gradually　liberalized　the　control　of　COVID-19,　which　has　caused　a　wide　discussion　in　society.　In　this　paper,　Natural　Language　Processing　(NLP)　algorithm[1]　has　been　used　to　analyze　online　comment　data　on　relevant　topics　on　the　SINA　Weibo　platform　from　November　2022　to　March　2023.　These　data　are　related　to　the　public＇s　liberal　views　on　COVID-19.　This　research　is　divided　into　three　steps:　(1)　Use　the　Octopus　Data　Collector　software　to　crawl　the　public　opinion　comment　data　in　the　SINA　Weibo　platform.　Then,　the　crawled　data　has　been　cleaned,　segmented,　and　manually　annotated.　At　the　same　time,　a　marked　data　set　has　been　obtained　from　the　Paddle　platform[2].　(2)　A　total　of　45,000　processed　training　data　has　been　input　into　the　Attention-Bi-LSTM　model　for　training,　and　the　optimal　model　has　been　selected　based　on　the　three　evaluation　indicators　of　Accuracy,　F1-score,　and　Recall　for　subsequent　sentimental　prediction.　(3)　Finally,　the　model　has　been　utilized　to　predict　the　remaining　24145　comment　data.　The　analysis　results　showed　that　after　the　announcement　of　the　release　of　COVID-19　Epidemic　prevention　and　Control　(COVID-19　EPC),　the　number　of　daily　relevant　comments　on　the　SINA　Weibo　platform　has　declined,　indicating　that　the　life　of　the　Chinese　public　has　begun　to　shift　toward　a　better　situation.　It　is　proven　that　it＇s　feasible　to　use　an　NLP　algorithm　to　predict　the　sentiment　of　the　online　topic　at　a　certain　time,　and　it　provides　a　potential　method　for　the　mining　of　the　Chinese　public＇s　online　public　opinion.　©　2023　IEEE.