In the paper, a problem of modelling size, structure and dynamics of a human capital is solved. The research object is an economic system. The subject is a practice of applying neural networks to socio-economic parameters modelling, specifically a human capital. The objective of the paper is to build an adapted neural network algorithm with the purpose of modelling the parameter being studied. Two human capital components are estimated; these are its quantitative and qualitative properties. The key element of a quantitative property (namely, the population reproduction) has a bearing on stability of a human capital development. The quantitative property is multifold: its aspects are healthcare, culture, education and science. To estimate a human capital structure, a population is being divided onto social clusters on the basis of these aspects. As a part of the study, it was found that such mathematical modelling instrument as neural networks is very suitable for conducting a cluster analysis of a given social system. Neural networks are effective means to solve poorly formalized problems; they are tolerant to frequent changes of an environment and can be used to process a vast set of contradictive or incomplete data. The data base comprises demographic data, volume of investments into qualitative human capital properties, and socio-economic development indicators of a given economic system. A gradation of demographic elements of the society based on physical condition and cultural and educational level is built, according to which a statistical data is gathered to solve the clusterization problem. A volume of investment into a human capital is defined by budgetary costs and private investments of the people. Modelling human capital investment dynamics is performed with neural networks being applied as well. The neural network model used herein is a multilayer perceptron with sigmoid logistic activation function. Neural network modelling of predicted values of investment volumes has proved its effectiveness. An estimation of a human capital for the period of 2000-2019, as well as its forecast for years 2020-2025, is exemplified by Russian economic system. Calculations showed that the indicator being studied has been demonstrating the biggest growth rates since 2013, with an ongoing growth to be expected. Evaluated results correlate with a Russian human capital index dynamic pattern, which is defined by UN specialists, qualitatively. A proposed method of a human capital prognosis and estimation can be used furthermore to compare and estimate socio-economic state of Russia’s regions.