This paper designs and proposes a constructive recommendation model framework composed of knowledge network module and constructive recommendation module, adding constructivist recommendation strategy to form a personalized knowledge recommendation model based on constructivist learning theory. Improve the insufficiency of matrix decomposition, design the user feature generation, information matching and generating recommendation three parts to constitute the big data recommendation module. Design the teaching quality assessment module and apply it to the assessment of teachers’ self-improvement ability development level. Taking the practice course as an example, we analyze the accuracy of the personalized knowledge recommendation model based on constructivist learning theory in recommending learning resources during teachers’ online learning process. Based on the evaluation results of the teaching quality assessment module, we found that there are significant differences in the self-improvement practice dimensions among teachers of different age groups. The scores of self-improvement awareness dimension, learning dimension, and practice dimension are good. The overall assessment result of teachers’ self-improvement is 77.54, which is a good overall level. Teaching practice using personalized intelligent knowledge recommendation model can assist teachers’ self-improvement and can be used as one of the paths for teachers’ professional development.
