This study proposes an AI-based recommendation system for tourist destinations that incorporates personalized criteria to generate a preference ranking tailored to individual tourists. The model was applied to the case of Varanasi, India, where four domain experts provided linguistic assessments of selected tourist sites. These qualitative inputs were transformed into triangular fuzzy numbers, and the Fuzzy TOPSIS method was employed to derive a robust ranking of destination preferences. To ensure the model’s reliability under uncertainty, a sensitivity analysis was conducted with respect to the fuzzification factor, revealing consistent ranking results and confirming the system’s stability. The proposed approach offers a improved, beneficial, and adaptive decision-support tool for the tourism sector, enabling tourists to plan or revise their itineraries according to personalized needs and preferences. Furthermore, the generated rankings can support policymakers, administrative bodies, and tour operators in evidence-based decision-making and strategic development of lesser-ranked destinations. By aligning individual tourist choices with policy objectives, this system contributes to more balanced and sustainable tourism growth.