At present, all the economic and social fields are rapidly evolving from digitization, networked to intelligentization. It is already a heated research topic how to apply the artificial intelligence technologies to improve the intelligent level of civil construction industry. The BIM technology is an important achievement in the digital and networked development of this field. When it enters the stage of deeper, popularizd and large-scaled application of BIM technology, how to get a multi-directional, objective and quantitative evaluation of the quality of BIM models automatically and intelligently becomes an important issue. This paper combines the research and developmentpractice of the BIMChecker in our research group, and elaborates how to apply the domain-specific language design, the natural language processing, the knowledge graph, the data mining, the machine learning and other important technologies in this filed. The techniques are applied to realize the understanding and representation of building codes, the extraction and organization of semantic information of BIM models, new knowledge discovery, and automatic identification of drawing symbols, etc. The effective applications of the BIMChecker on typical projects are also demonstrated. This paper provides reference for exploring the effective combination and application of artificial intelligence technologies in the civil construction industry.