FAQ are often built up by the customer service department in enterprises as a knowledge base to facilitate the reusability of solution for frequent encountered problems. However, subjective judgments might cause inconsistent classification, which lead to inappropriate response to users' inquiries, and also retard the searching efficiency. This research is taking on-line service FAQ of IT Company S as a model to study the FAQ Automatic Categorizing System, which is expected to offer a better searching quality to respond to users’ inquiries in the future. First of all, the domain ontology for the case S is designed for the research, then, the classification method and rule is built to enable the automatic creation of classes by the system. After that, the System Development Research Methodology is used to develop the FAQ Automatic Categorizing System. Finally, interviews and surveys are conducted to explore the users' satisfaction with the system, evaluate the feasibility, and investigate the searching efficiency of the system. The result of this research showed that FAQ Automatic Categorizing System is capable of creating the classes that matches the FAQ data automatically with timely adjustment by manual to show better performance. After the investigation, the result shows that it has higher efficiency to search the FAQ using the question attribute than using the software name. Although the original system S company using has the same satisfaction to that of the FAQ Automatic Categorizing System, but the FAQ Automatic Categorizing System stands out when doing the FAQ search. It was expected to regard FAQ Automatic Categorizing System as a model for the future research and practice.