The intelligent education recommendation system can recommend knowledge suitable for students' personal learning. However, the traditional recommendation algorithm has generality problems, which lead to poor knowledge recommendation effects. In order to improve the performance of the education recommendation system, based on the machine learning algorithm, this paper combines the knowledge graph algorithm to improve the recommendation algorithm and decomposes the matrix with a higher dimension into several matrices with relatively small dimensions through matrix transformation. Moreover, this paper conducts in-depth mining of the potential attributes of users and items and improves the matrix decomposition formula based on knowledge recommendation requirements. In addition, this paper constructs the framework of the intelligent education recommendation system with IoT networks based on the analysis of functional requirements. Finally, this paper designs experiments to verify and analyze the model from the perspective of model performance and user satisfaction. The research results show that the algorithm model constructed in this paper is effective.