With the development of financial institutions, this application software and related information technologies are used not only by specialists, but also by ordinary citizens to solve tasks that a few years ago seemed to be within the competence of only mathematicians specializing, for example, in building forecasting models. It can be noted that the collaboration of IT with application software, as well as with the mathematical apparatus most typical for forecasting tasks, gives good results. In particular, this applies to the Central Bank market. The study is devoted to the problem of approaches to the selection of methods and strategies for analysis and forecasting of the central bank markets, which is an urgent issue today. Far from all possible methods and strategies have sufficient coverage in the scientific information space, which prompts the need to analyze and systematize already existing information in this field. Accordingly, basically. the purpose of the study is to analyze and systematize the theoretical foundations of existing approaches to forecasting the CCV market. An analysis and systematization of the theoretical foundations of existing approaches to forecasting the CCV market was carried out. Generalized advantages and disadvantages of structural methods and models used for making market forecasts were outlined. A comparative analysis of ANN models was carried out in terms of their use for market analysis tasks. Among the analyzed ANN models are the following: CNN-2l, CNN-3l, LSTM, sLSTM, BiLSTM, GRU, CLSTM, MLP and RFBNN. The analysis and testing of existing models provided results that provide a wide scope for further research and study.