Today, people are more focused on personal experience and individual needs when travelling, such as expecting a low cost, convenient and comfortable travel experience during their trip, as well as having different visiting needs such as historical and cultural, natural landscapes. This paper proposes a method for recommending tourist routes based on Random Forest and information enthalpy evaluation system. Taking Hohhot city as an example, data on the location, ticket price, rating and playing time of attractions in the area are obtained, and a comprehensive evaluation model considering the importance of influencing factors is constructed from three factors: attractions, traffic and tourists using Random Forest, and the weights of the comprehensive evaluation model are estimated using the information entropy assignment method to screen out high-quality tourist attractions. Finally, based on the different needs of different groups of tourists, an attraction selection model with the number of attractions and the overall rating of attractions as the objective function is established and solved according to the geometric weighting method. Compared with the traditional travel route recommendation method, the proposed method takes into account the user's needs and allows for as many high quality attractions to be visited with limited time and expenditure.