Information technologies penetrate into all spheres of human life and the activity of companies. The sphere of tourism did not escape their influence. Tourism is a leisure activity that involves complex decision-making processes, which is why the development of recommendation systems that will help facilitate these processes is relevant. It is for this purpose that the method of selecting tourist routes using machine learning methods was created in the work. In the course of the research, the field of tourism was analyzed, the peculiarities of recommendation systems were established. The levels of the proposed methodology are described, indicating the features of the backend and frontend. A characteristic feature of the proposed methodology is the use of machine learning classifiers, such as decision trees, support vector method, and multilayer perceptron, in the model of optimal selection. It is this combination of methods that will allow you to choose the best route from the proposed ones. They are used to distinguish specific destinations within each data set. To make the complex model usable and interpret its results for the tourist, the decision tree models are transformed into decision rules, and then the information is passed to the user interface control module. In order to obtain a high-quality forecast, the methods of machine learning were considered in detail, and their features were determined. It has been found that the classification performance can be improved by using a combination of classifiers and methods such as decision trees, support vector method and multilayer perceptron. This combination of algorithms will allow to predict the optimal tourist route. Both technical and practical aspects, including data sparsity, scalability, transparency, system accuracy, usability, and user acceptability, were considered to evaluate the effectiveness of the proposed methodology.