Sentiment analysis of visitors to the tourist destinations of Borobudur Temple in Indonesia needs to be done to determine the expected product and service preferences. In addition, sentiment analysis is also helpful for managers to adjust the needs of tourists to the infrastructure provided in the tourist destination area. The classification method used in the sentiment analysis is the Naïve Bayes Classifier (NBC) against 3850 visitor reviews at Borobudur Temple. Review data is pulled from Tripadvisor web pages filtered by language, review time, and travel characteristics to analyze foreign traveler preferences comprehensively. This research stage is divided into three parts: data preparation, data processing, sentiment analysis, and algorithm performance evaluation. In addition, SMOTE Upsampling is used to balance data. The results of implementing the Naïve Bayes Classifier (NBC) classification method obtained an accuracy value of 96.36%, a precision value of 93.23%, and a recall value of 100% with an Area Under Curve (AUC) value of 0.714. In addition, the results of ranking five famous words from the review data show that there are highlights of the physical condition of the temple, scenery, and tourist visit activities at Borobudur Temple, where the four most famous words in visitor reviews are the “temple,” “visit,” “Borobudur,” “sunrise” and “place.”