Abstract: Participating in the global AI innovation chain competition actively and seizing the dominant position of the innovation chain are of great significance for China to accelerate its construction into an innovative country and a world science and technology power. Based on the existing theoretical research of innovation chain, combined with the technical and economic characteristics and industrial characteristics of AI, an AI innovation chain model is proposed, and the four major innovation links of knowledge, technology, products and services, and the four characteristics of diversity, collaboration, circulation and development are sorted out. Using the model structure, the analysis shows that the current global AI innovation chain has basically formed a pattern dominated by China and US, and led by East Asia, North America and Western Europe. Compared with other leading countries, China still has many deficiencies, including the lack of theoretical discourse power in knowledge innovation, the lack of original subversion in technological innovation, the lack of the innovation support chain construction, and the imbalance in the industrial chain, supply chain and innovation chain. To enhance the competitiveness of AI innovation chain continuously, China should promote the collaborative innovation of basic theories and key common technologies, promote the integrated development of industrial chains and innovation chains of AI with other related industries, improve technological governance capabilities, attract high-end talents to participate, and expand international cooperation.