【Purpose/significance】In the information age, online public opinion is spreading wantonly, and the smart library bears the

responsibilities of information service and knowledge organization and management, so it is important to build an online public opinion service system to improve users' public opinion information literacy and comprehensive intelligence capability and cooperate with gov⁃ernment departments in public opinion management.【Method/process】Following the user-centered service principle, on the basis of analyzing the subjects of public opinion services of the smart library and their information needs, this paper constructed an online pub⁃lic opinion service system from five levels of demand perception, data perception, context perception, situational awareness and intelli⁃gence service by using intelligence perception theory. Besides, the guarantee mechanisms are proposed for the possible problems in the process of service implementation.【Result/conclusion】The constructed online public opinion service system in the smart library can realize public opinion intelligence perception at the levels of demand, data, context and situation, and provide users with smart, personalized and diversified public opinion knowledge services.【Innovation/limitation】This study can expand and deepen the theory and service model of the smart library, and highlight the value of public opinion information services and digital inclusion capability of the smart library.