In modern society, with the profound integration of intelligent technologies, social network platforms have gradually become the primary medium through which the public constructs spatial cognition. This paper delineates the relationship between User-Generated Content (UGC) data and tourists’ perceptions employing the ROST-CM6 text analysis tool to systematically examine the contents of online tourism reviews and travelogues of typical traditional villages within the Li River basin. By integrating online text analysis and affective tendency assessment methods, the study quantitatively evaluates the significance and expressiveness of tourism perception elements using the Importance-Performance Analysis (IPA) evaluation model. Findings indicate that the core components of tourism image perception of traditional villages in the Li River Basin encompass five aspects: natural landscape, humanistic landscape, facilities and services, cultural experience, and tourism evaluation perception. Additionally, four issues are identified: imbalance in resource utilization and misalignment, insufficient supply of facilities and services, conflict between commercialization and originality, and ‘superficiality of cultural experiences.. The paper proposes an optimization pathway for traditional villages in the Li River Basin, comprising “spatial reconstruction, cultural empowerment, and service improvement”, aimed at fostering the preservation and inheritance of traditional villages and offering a reference for analogous regions.