With the development of emerging social media like Microblog, Douyin and Tieba, an increasing number of users have been drawn to these platforms to post and acquire information, leading to the accumulation of massive public opinion data. In order to promptly monitor public opinion movements and better safeguard the safe operation of the Internet and network security, a SnowNLP-based network public opinion analysis system was designed to address real-time Microblog data. The system was comprised of the acquisition, analysis and visualization of public opinion data, which is capable of textual mining of Microblog data, emotional analysis of network public opinion data, match of public opinion data with keywords, and visualization of results of emotional analysis of Microblog content, user level and word segmentation. The results of the experimental test show that the features of the system function normally and in the meantime the feasibility and effectiveness of the design solution are validated. The system is of great applicative value in the field of network public opinion monitoring.