A learning resource recommendation method based on fuzzy logic (RRMF) was proposed to solve the ambiguity problem of categorizing learning resources and determining learners' knowledge levels in the learning resource recommendation process and to improve the accuracy of recommending learning resources. First, learning resources were modeled using fuzzy set theory. The learning resources were divided in accordance with knowledge points, and the association of each learning resource to different knowledge points was calculated. One learning resource is associated with multiple knowledge points at different degrees. Second, the learners were modeled through fuzzy cognitive diagnosis method by determining the knowledge level of learners. On this basis, a learning resource recommendation model was constructed. Scores of candidate learning resources of the target learners were predicted, and learning resources with scores that satisfy the value range were recommended to learners. Finally, the proposed RRMF was verified through a test on real datasets FrcSub, Math1, and Math2. Experimental results demonstrate that the mean absolute errors of the proposed RRMF on different datasets are 0.3045, 0.2944, and 0.2817. The proposed RRMF is remarkably superior to the recommendations based on cognitive diagnosis, probabilistic matrix factorization, and collaborative filtering. Moreover, the proposed RRMF increases the accuracy of recommendation. This finding confirms that the proposed RRMF can solve the fuzziness problem in the learning resource recommendation process effectively and increase the accuracy of the learning resource recommendation. The conclusions in this study provide novel methods for designing learning resource recommendation services. [ABSTRACT FROM AUTHOR]