This article explores modern methods, models, and technologies for forecasting tax revenues. It discusses the stages of neural network modeling in forecasting tax revenues. Neural network technology enables rapid assessment of all factors affecting tax revenues and enhances the system’s response to these influences. As a result, it expands the ability to accurately forecast tax revenues. Neural networks imitate human brain activity and are systems or programs built on artificial intelligence. The scientific novelty of this study is that it overcomes the limitations of traditional statistical and mathematical methods in the process of forecasting tax revenues and creates the opportunity to model complex and nonlinear economic relationships. The approach developed in the article based on neural networks increases the accuracy and stability of forecasts by taking into account macroeconomic, fiscal and external factors affecting tax revenues. In addition, the proposed model allows for the effective processing of large amounts of real-time data generated in the process of digitization of the tax system and is of significant scientific and practical importance for practical application in public finance management. As a result, this study reveals the scientific basis for using neural networks in the process of forecasting tax revenues and serves as an effective tool for sustainable management of public finances.