Generative AI technology’s fast expansion is driving increasing application in visual creation. This work presents an artificial intelligence generative visual creativity system (KRGVS) with knowledge retrieval to automate user intent understanding and high-quality visual content creation. Its key uniqueness is the ability of the KRGVS system to extract deeper purpose from confused user inputs, call suitable knowledge resources, and produce innovative and valuable visual creations based on a semantically enriched generative model. Experimental results imply that the KRGVS system can use cross-modally knowledge to improve visual content creativity, logic, and explanatory power, generation quality, and user customisation. This paper provides a new data-driven and intelligent aid direction for the creative industry and supports the change of AI from a content generating tool to a human-machine collaborative creativity partner.