China＇s　five　thousand　years　of　rich　history　have　nurtured　many　historical　and　cultural　districts　with　profound　cultural　heritage.In　recent　years,　these　districts　have　become　popular　tourist　destinations　for　both　domestic　and　international　travelers.This　paper,　using　Fujian　Province　as　a　case　study,　conducts　sentiment　analysis　and　LDA　topic　modeling　on　online　reviews　to　explore　the　factors　influencing　tourist　satisfaction　with　historical　and　cultural　districts　and　proposes　suggestions　for　improving　tourist　satisfaction.Additionally,　addressing　the　current　issue　of　low　accuracy　in　sentiment　classification　of　online　reviews,　this　paper　introduces　an　improved　model　based　on　RoBERTa_wwm　word　embeddings　and　Bidirectional　Gated　Recurrent　Unit　(BiGRU),　incorporating　an　attention　mechanism　to　highlight　the　emotional　words　in　the　text　that　have　a　greater　impact　on　classification　results,　thereby　enhancing　the　accuracy　of　sentiment　classification.Experimental　results　demonstrate　that　the　model　performs　exceptionally　well　across　various　datasets.Moreover,　LDA　topic　modeling　reveals　that　tourist　reviews　are　primarily　focused　on　local　characteristics,　with　particular　attention　given　to　aspects　such　as　the　history　and　culture,　local　specialties,　snacks,　and　architecture　of　historical　and　cultural　districts.　©　2024　Copyright　held　by　the　owner/author(s).　Publication　rights　licensed　to　ACM.