At present, the web text has become an important source to help understand tourists' experience and perceptions, which has attracted the attention of academia. Web crawler technology is used to crawl online tourist comments about Sanhe Ancient Town, which are processed by ROST Content Mining software to filter out high - frequency feature words. By using web text analysis method, part-of-speech and frequency analysis of high - frequency words, semantic network analysis, and sentiment analysis are conducted. The results show that the tourists' perceptions of Sanhe Ancient Town mainly involve the four dimensions of location, attractions, infrastructure, and tourism environment. Tourists generally have positive attitudes towards Sanhe Ancient Town, and they show overall satisfaction with their travel experience. It is suggested that Sanhe Ancient Town should improve tourists' experience and perceptions from the aspects of maintaining regional advantages, building a characteristic cultural town, "revitalizing" culture, and beautifying the environment.