applications designed to adjust prices depending on fluctuations of supply and demand. Even though the mentioned element seems to be effective in terms of the allocation of vehicles and management of the process of providing a certain level of service quality, it is rather obscure, which causes concerns regarding its efficiency and fairness. Therefore, this paper analyzes the peculiarities of the operation of the surge pricing mechanism during periods of high demand relying on the data collected from a ride-hailing application.As part of this project, researchers developed predictive models utilizing histogram gradient boosting, XGBoost, LSTM, and Prophet methods. These models are characterized by their ability to analyze relationships between different variables as well as to establish dependencies based on temporal characteristics of the process, which allows for creating accurate predictions. The results of this research indicate that the process of dynamic pricing has many complex aspects that cannot be identified by conventional means.