The current application of intelligent technology in the field of education is becoming more and more widespread, especially in personalized English teaching, which shows enormous potential and value. After combining the concepts of personalized learning paths with algorithm optimization, the study gathers online learning behavior data from university students as an example and investigates the factors that influence English learning through correlation and regression analysis. Afterward, the K-mean clustering algorithm is optimized by improving the method of determining the number of clusters, and the learner analysis is completed based on the learning behavior data to divide the sample students into clusters. On this basis, the personalized learning path under the university English teaching mode is designed and applied to the teaching practice to explore the students’ learning effects after the path is applied. The sample students can be divided into three categories: low-invested, medium-invested, and high-invested, which account for 10.19%, 25%, and 64.81%, respectively. The English scores of the students who applied the personalized learning path were significantly improved, with 7.037 and 6.985 points for freshmen and sophomores, respectively, and there was also a substantial improvement in English learning attitudes and learning strategies, which verified the practical effect of the designed personalized English learning path.