Abstract: In order to study consumer's preference for pipe tobacco products, a web crawler based on Python programming language was used to obtain the pipe tobacco product information and network evaluation data. The consumers' preference for brand and formula type of pipe tobacco products were analyzed based on the data. In addition, consumer preference for flavor types and formula compositions of pipe tobacco were also analyzed by word frequency analysis method and association rule mining method. The results showed that: 1) Pipe tobacco products from large tobacco companies had higher market shares. 2) Consumers presented obvious preference for flavor types. The pipe tobacco products with original taste (None) were favored. In condiment pipe tobacco products, Vanilla, Rum, Cocoa/Chocolate and Citrus-flavored pipe tobacco products were popular. The pipe tobacco products of Aromatic, English and Straight Virginia were more popular and had higher market acceptance. 3) From the formula compositions, tobacco leaf materials, such as Virginia, Latakia, Burley and Oriental/Turkish were commonly used. The original types of pipe tobacco products formulated with the above leaf materials were most popular in the market. The direction for developing pipe tobacco products could be obtained through the analysis of the market preference on the basis of internet data.