In the age of digital intelligence, and based on the big data technology, the work aims to mine online user reviews of Cai's Ancient Residence, discover its existing problems and propose effective design strategies to verify their feasibility. Firstly, the work mined the user comment data about Cai's Ancient Residence on popular travel platforms and social media on the Internet, analyzed and summarized user needs, and determined the development of a brand cultural tourism IP image that could inherit the architectural genes of Cai's Ancient Residence, and used it as an attraction. The objects were placed in the virtual navigation program and public display space to enhance the attraction of Cai's Ancient Residence Scenic Spot; Secondly, based on the architectural meme theory, an architectural gene map of Cai's Ancient Residence was constructed, and their architectural form genes, spatial genes and thinking genes were mapped respectively. Then, the S-O-R theoretical model was used for experimental evaluation, and architectural factors were selected for design based on user response time; Finally, with the help of artificial intelligence generation tools, the extracted architectural factors were applied to the design of IP image plans. The optimal solution among the intelligently generated images was selected, and three-dimensional modeling was completed manually and applied to the online and offline tourism promotion scenarios of Cai's Ancient Residence. This study proposes an innovative framework that combines big data mining, building meme theory and artificial intelligence auxiliary design. It has important theoretical research significance and practical application value for inheritance and communication of cultural heritage genes.