With the increasing development of information technology in China, the application value of social media has been further highlighted. Compared with traditional media, the interactivity of social media is more in line with the requirements of students’ academic communication, which is conducive to students’ discussion of their own views and enhancement of their learning effect. This paper proposes the realization path of social media in the construction of the English classroom, through the network teaching, interactive teaching and other methods, to carry out the research on the transformation of English education to digitalization. Using two-branch network DNN-CBLM, a classification prediction model of English performance is constructed based on students’ online learning behavior. BP neural network was introduced to establish the factors influencing the quality of English teaching and to realize the effective evaluation of English teaching quality based on the objective attributes of English teaching quality evaluation.91.348% of the students were satisfied with the use of social media for their English learning, and compared with the traditional media teaching class, the English performance of the experimental class of social media teaching was improved by more than 5 points, and the mean value was 6.337% higher than that of the control class. Further analyzed using the independent sample t-test, the sig value of the mean equation of the two groups of experimental subjects is 0.004, which is lower than 0.05, indicating that the students in the experimental class have a higher understanding of the knowledge points and a better learning effect under the social media-based teaching environment.
