This paper focuses on constructing a regional digital educational resource sharing platform. It applies optimized ant colony clustering technology to support the platform’s personalized service to achieve accurate distribution of educational resources. The research subjects were undergraduate and graduate students majoring in mathematics in teacher training, and by comparing their teaching abilities before and after using and not using the platform, it was found that the average scores of students using the platform were significantly higher than those of students who did not use the platform on each evaluation factor. Undergraduate students’ information literacy competence improved with the academic year, with first-year students’ scores between 20 and 50, while senior students’ scores between 70 and 100. At the graduate level, the difference between first-year graduate students and junior undergraduates was significant, but the difference with senior undergraduates was slight (p-values of 0.042 and 0.053, respectively). The digital education resource sharing platform effectively improved the teaching ability of teacher students. However, most of the students lacked experience in informatization practice activities.