With the rapid development of educational technology, personalized learning serves as one of the important means to improve the quality and efficiency of education. However, traditional teaching methods often neglect individual differences of students, resulting in unsatisfactory learning effects. To solve the current problem of low learning efficiency of learners in the field of education, a personalized test question recommendation method based on a knowledge graph is proposed. First, a comprehensive and detailed knowledge map of knowledge points is constructed by integrating expert knowledge and fine modeling of knowledge points. Then, combining the dependencies between knowledge points and learners' knowledge mastery, we designed a novel test question recommendation algorithm, which fully considered the contribution value of each knowledge point. During the implementation of the algorithm, we effectively predicted the recommendation degree of the alternate test questions by updating the learner's knowledge point loss matrix and accurately recommended the test questions with a high recommendation to learners. Through experimental validation, the results show that the proposed algorithm can significantly improve the learning efficiency of learners and provide an effective solution for personalized learning.