Abstract: 【Purpose/significance】This paper explores the transfer and evolution rules of knowledge acquisition behavior of research⁃ ers under the background of generative artificial intelligence, in order to provide theoretical references for in-depth understanding of knowledge acquisition behavior patterns of researchers and guide researchers to rationally use generative artificial intelligence knowl⁃ edge acquisition【. Method/process】Based on the finite rationality hypothesis, from the perspective of perceived benefits and perceived costs, this paper constructs an evolutionary game model between researchers using generative artificial intelligence knowledge acquisi⁃ tion and traditional search behavior knowledge acquisition, analyzes the evolution process of researchers from traditional search behav⁃ ior knowledge acquisition to generative artificial intelligence knowledge acquisition behavior, and uses the simulation model to explore the effects of various factors on the behavior evolution.【Result/conclusion】The research shows that generative AI knowledge gain, knowledge search cost and emotional support benefit positively affect the knowledge acquisition of researchers using generative AI, while knowledge search gain, generative AI application cost, information risk cost and other factors negatively affect the knowledge ac⁃ quisition of researchers using generative AI【. Innovation/limitation】In this paper, the two knowledge acquisition methods of research⁃ ers are regarded as game processes, and the influence of each factor on the behavior choice of researchers is explored from the perspec⁃ tive of perceived benefits and costs. In the future, other factors such as individual cognitive bias and technical limitations can be fur⁃ ther considered in the study.