Abstract. The paper deals with the research of the process of economic growth as a component of the economic development of countries and aims at developing a neural network model directed at improving the modeling of economic growth, its stabilization or recovery after the impact of globalization integration processes, crisis phenomena. During the analysis, the authors found out that the indicators used to create a model of economic growth by the countries of the world do not have a close correlation and reflect different conditions of their functioning. It is determined that in order to achieve the set goal, it is advisable to apply neural networks, which provide a possibility to build a forecast system of economic growth with greater accuracy. In the process of the analysis, the data of 77 countries of the world according to the indicator of economic growth were used, the level of economic growth of the countries was assessed, the most accurate neural network model and the optimal network architecture were determined. The authors of the paper solve the problem of approximation of experimental data using multilayer perceptron-type models and network models with radial basis functions. The dependent variable in the model is denoted by the level of economic growth, and the independent variables are the level of gross accumulation, the level of gross savings, the level of export of goods and services, the level of import of goods and services, the level of current health care expenses. The volumes of training, test and control samples, the developed neural network models and the obtained results of economic growth modeling are presented graphically in the paper. The neural network model developed by the authors is sufficiently adequate, which is confirmed by the volume of processed data and obtained results. The neural network model of economic growth is suitable for further use in the process of its forecasting in various countries of the world.