As an important means to improve the efficiency of English teaching, the current method of student stratification management still has shortcomings such as being too broad and not fully considering the students’ situation. This paper organizes the relevant concepts of Bayesian network, and on the basis of this theory, proposes Bayesian network prediction as a prediction method of students’ performance. At the same time, in order to analyze and predict students’ performance more scientifically and objectively, the simple Bayesian classification method is introduced. Combining the traditional weighting method and informatics method to calculate the weight value, the premise of extending the conditional independence of the simple Bayesian classifier. Based on the prediction of students’ performance, the layered teaching concept is used as a framework to design the English layered teaching model and the learner’s layering method. The stratified management of students is realized by adopting corresponding teaching methods based on students’ English achievement characteristics. The tiered management method was recognized by 90.00% and above of the students in the practical application.