Building information modeling (BIM), characterized by multi-dimensional modeling and multi-source data integration, is becoming a key technology to promote the innovation and digital transformation of the highway design industry. Taking BIM as the design deliverable is an inevitable trend in the future development of highway engineering design. However, the current highway BIM review is mainly conducted manually, which brings about problems such as low efficiency, susceptibility to errors, and high subjectivity. This makes it difficult to adapt to the review requirements of the three-dimensional (3D) digital design mode. To this end, this study proposed an automatic review method of highway BIM based on a knowledge graph. By constructing the knowledge graph covering multi-dimensional knowledge such as highway design standard specifications, a semantic database, and meta-structures in the field of highway engineering, natural language processing (NLP) technology was used to structurally process the review rules of design standard specifications and provisions. Then, by taking a component entity of industry foundation classes (IFC) as the object, the Cypher query language was used to realize the review of attribute information completeness, data accuracy, and design compliance of highway BIM components. The results show that the graph database based on the knowledge graph technology can provide technical methods for the review of highway BIM design results, further significantly improving the quality and review efficiency of highway 3D design results.