In this paper, the clustering algorithm of big data fuzzy K-means is used to research and analyze the teaching management system. Since the values of the indicator variables of bad units are usually significantly different in smooth areas and discontinuous areas, we cluster the indicators of local areas through K-means clustering, so that only good units or bad units are included in each class. The server is deployed on the campus and is networked with the teaching computers of each classroom. Teachers log in to the class rate management system before or during class, log in to their respective users, select the corresponding class, and make corresponding attendance records according to the attendance of students in the class. The system automatically counts the number of expected arrivals, attendance, late arrivals, leave requests, and absenteeism for the class. Teachers can also fill in information feedback for this class: a list of absenteeism, classroom discipline, student learning status, teaching suggestions, and teaching testimonials. The development environment built by this system is a combination of PHP + MYSQL, which reasonably planned according to the overall needs. The development mode adopts the top-down model, which runs through the whole process from system development to testing to application. The most used floating point number encoding is the decimal floating point number encoding. The relationship of each module of the virtual reality teaching management system is sorted out, the overall framework of the virtual reality teaching management system is established, and the design of the business flow diagram of the management system is completed. This paper also uses Oracle database technology to further explain the design process of designing the system database. Finally, combined with the system design and implementation, the system test is completed from two aspects: system function test and performance test.