This paper primarily designs and implements a collaborative filtering recommendation algorithm, which is based on user, content, and student profiles. It constructs the user’s learning style or learning interest by analyzing the user’s historical behaviors, and then generates recommendation results by searching for the required information in the information resource database. Through the learner’s basic personal information and behavioral data, the learner is profiled, and the user profile is constructed to achieve personalized teaching in the English classroom, thus realizing the personalized classroom teaching mode combined with the recommendation algorithm. The results indicate that the average heat value of text-based English teaching resources in the cloud classroom is 0.220, higher than that of video-based resources. In terms of time commitment, most students have low levels of commitment to online learning, with all of their online learning hours being under 150 hours. The personalized classroom teaching mode can significantly improve students’ overall performance compared to ordinary teaching methods (p<0.01).