The invention relates to an Agent-based consumer preference analysis system and relates to the field of consumer purchase decision environment simulation. The system comprises a data visualization module used for collecting consumer preference data, mapping the collected consumer preference data into recognizable graphs, images, videos or animations, and displaying the recognizable graphs, images,videos or animations to users, and an Agent module which is used for quantifying consumer preferences through a utility function and analyzing consumer preference behaviors in combination with the acquired consumer preference data. According to the system, the knowledge graph of consumer purchase is drawn by utilizing CiteSpace software, the information panorama of consumer purchase behaviors isvisually displayed, and key influence factors in the consumer purchase behaviors are identified; heterogeneity of system members is reflected on an individual level of consumers by utilizing a bottom-up modeling mode based on the Agent, shopping behaviors of the consumers are simulated by comprehensively considering factors influencing the purchasing behaviors of the consumers, and decision of thepurchasing behaviors of the consumers is analyzed and predicted through the emergence phenomenon of the model.