In this study, we proposed the “Prediction Photo System” which leverages Digital Twins and Generative AI technologies to predict and present photos that sightseers might take a few minutes later and examined how this influences their destination choices and overall sightseeing experiences. In a field experiment, detailed records were kept of system usage, participant behavior, and feedback from a total of six participants. The experiment demonstrated that the system effectively influenced sightseers’ behavior, as almost all participants exhibited behavior guided by their own “prediction photos.”. However, each participant experienced the system differently, highlighting its potential for personalized sightseeing experiences. Additionally, several system improvements, such as time settings, location information, and interface design, were identified for future enhancements.