Purpose of this study was to develop the main components of a model of an adaptive information system for predicting crop productivity. Methods. To conduct research on the establishment of the basic structural elements of an adaptive information model for predicting the productivity of basic crops used the method of constructing dynamic models. Results. A detailed analysis of conceptual approaches to the construction of mathematical agricultural models is carried out and the main advantages and disadvantages of modern analogues are established. It is determined that the adaptive information model is based solely on the needs of the plant and actually on the need to provide these needs with available resources in order to obtain consistently high yields with high quality indicators. The hardware and software complex must have a feedback relationship between its basic structural elements, because it significantly improves the accuracy of predicting plant productivity. Data based on the operation of certain mechanisms or indicators of weather conditions and their forecasts are used for decision making, however, if they are substantially changed, decisions about individual technology elements are reviewed. The software should be related to the economic part and should take into account market conditions and forecast data when making recommendations. In the case of low purchase prices for products, we recommend that certain agrotechnical operations (say vegetation feeding) be applied or not, in the case of significant change in growing conditions - when the application of these agro-measures will be ineffective due to the negative effects of drought, etc. Conclusions. Adaptive information system for forecasting productivity in the technological process of growing crops is formed on the basis of a model consisting of three modules of characteristics – the resultant and two components. At each subsequent stage of implementation of the model, the resulting feature becomes component, with the maximum contribution to the resulting feature of the next module.