This study aims to discover and analyze the most valued factors in rural tourism. An analysis of a sample with 9,939 comments obtained from the booking platform Booking.com on stays in rural tourism establishments located in Portugal is presented. Using text mining techniques and topic modeling in order to find the most valued factors, word clouds were obtained that allow a quick and intuitive analysis of the data. The LDA algorithm allowed to extract ten topics coherent with the word clouds. The results point out that breakfast, the friendliness of the staff, the hosts and the outdoor space are the most relevant factors for guests. These results have implications for the management of rural tourism units as they identify the factors to be taken into account by those in charge of these units.