With the continuous improvement of the national education system, the Internet brings new opportunities for educational development but also brings specific challenges to the educational management of colleges and universities. This paper aims to analyze the current situation of education management in colleges and universities and construct an innovation model for education management clustering based on PDCA, which will be explored for its development value and power. And then, using the partial least squares regression algorithm and its modeling steps, to construct the educational management clustering regression model. The influencing factors of the educational management clustering model are analyzed with 12 “double first-class” colleges and universities in five provinces in Northwest China as research objects. The results show that the average comprehensive efficiency and pure technical efficiency of the universities in Northwest China range from 0.6585 to 0.9987 and from 0.773 to 0.994, respectively, indicating that the efficiency of higher education resource allocation in Northwest China is constantly improving. Compared with the traditional education management model, the colleges and universities applying the education management clustering model constructed in this paper in 2022–2023 show a scale-invariant trend, indicating that the colleges and universities have sufficiently invested in education resources and realized the scale efficiency. Educational management clustering can improve the efficiency of school educational resource allocation, thus improving the level of educational management and providing a theoretical basis for the study of innovative strategies in educational management.