A machine learning model for predicting canceled reservations based on the weather forecast from this project as well as an application interface have been developed for the purpose of maximizing revenue in revenue management in the tourism and hospitality sector. The web application is used to display the results of the model and the basic operation on a particular set of data used in the work, while the collection, analysis, processing and they periodically perform data processing in the background and thus exclude the possibility of error by the user. Application endpoints have been developed so that the machine learning model can be easily integrated into any system, whether it is a central reservation system as an input data channel or another database as a repository for all data processing and analysis results. Compared to other, currently available software solutions used in this business domain, its advantage is that its operation does not require a larger amount of resources and there is no need to install additional support structures. To make the solution in the form of a web application, an additional bonus is that only an internet browser is enough to use it. Related software solutions, developed by larger companies, process a much larger number of input features, ie they take a much broader context in optimizing revenue management. Seen from that side, therein lies the biggest limitation of the current model as it is narrowly specified for forecasting cancellations based solely on the weather forecast. But this fact can certainly serve as a good guide in further model development.