Many hotel recommendation systems are very useful due to the fact that many travelers and tourists need them. They can act as filters for information by providing suggestions for users to choose a satisfactory hotel from among many options. However, there are some chal- lenges in most of the hotel recommendation studies which may not be effective enough to make good predictions. Travelers and tourists are usually interested in hotels’ features, numerical values, and proximity to important areas. Therefore, this research aims to suggest the use of the k-means machine-learning algorithm to recommend hotels in several cases. First, we use hotel data combined with the history of a user to improve and make it more accurate, then we combine them to make a recommendation. Secondly, we provide a recommendation through hotels rating and cities. We used hotel data in Saudi Arabia, especially since the number of people coming to it is increasing due to the fact that it has become a tourist area.