The study of tourism destination images is of great significance in the tourism discipline.Tourism user-generated content (UGC), i.e., the feedback on tourism websites, provides rich information for constructing a destination image.However, it is difficult for tourism researchers to obtain a relatively complete and intuitive destination image due to the unintuitive destination image display, the significant variance in departure time and data length, and the destination type in UGC.We propose TDIVis, a carefully designed visual analytics system, aimed at obtaining a relatively comprehensive destination image.Specifically, a keyword-based sentiment visualization method is proposed to associate the cognitive image with the emotional image, and by this method, both time evolution analysis and classification analysis are considered;a multi-attribute association double sequence visualization method is proposed to associate two different types of text sequences and provide a dynamic visual encoding interaction method for the multi-attribute characteristics of sequences.The effectiveness and usability of TDIVis are demonstrated through four cases and a user study.