In the era of artificial intelligence, online learning of English courses in colleges and universities has gradually become one of the mainstream learning modes. Based on the traditional teaching methods, this paper carries out the research on the optimization of English teaching path in colleges and universities. A micro-learning unit clustering model is constructed with four modules: data preprocessing, learning pattern mining, learning path diagram construction and micro-learning unit clustering. The model analyzes the learning state of learners through sequence pattern mining technology, and conducts orderly planning of learning resources based on learners’ characteristics. On this basis, this paper defines the online learning path planning problem and online learning path planning according to the continuity characteristics of learning knowledge points, and constructs the online learning path planning model. At the same time, the dynamic planning algorithm is selected to carry out the optimization of path planning. Based on the learning status of different learners, the optimal online learning path is planned to realize the optimization of English teaching path. Compared with similar classical algorithms, the online learning path planning model has the highest matching degree of 0.8 between the planned paths and the learning states of users under different learning resources conditions, which verifies the superiority of this paper’s model in the optimization of English teaching paths in colleges and universities.