Abstract: Existing models do not impose any constraints when using the knowledge graph embedding methods, which leads to noise permeating the underlying information along with the embedding. Meta-path-based recommendation algorithms only adopt a uniform weight assignment strategy without considering the nuances among meta-paths. Traditional models usually suffer from the problems of data sparsity and cold start. To address these issues, a knowledge graph embedded meta-path recommendation algorithm based on heterogeneous attention network (MRHAN) is proposed. By defining meta-paths to capture the complex semantic information between different types of entities and relationships, the proposed method can better utilize the rich heterogeneous information to alleviate the problems of data sparsity and cold start. Using knowledge graph embedding methods, a constraint approach based on node relevance is employed to model semantically related higher-order entities and relationships into unique meta-paths. A hierarchical attention network is used to model node preferences for different neighbors and different meta-paths, learning the weights of nodes for different neighbors and the weights of meta-paths for different recommendation tasks. Experimental results show that the proposed algorithm can effectively learn the representation of knowledge graphs, as well as the importance of node neighbors and meta-paths, and further alleviate the problems of data sparsity and cold-start.