GOAL: This paper is devoted to improvement the technical condition assessment system of transformer equipment by machine learning. Today, due to the emergence of a large number of methods and systems for assessing the technical condition, the issues of analysis and forecast the technical condition of transformer equipment are quite acute, because in most systems there is no sensitivity to various scheme-mode parameters of the interconnected power system (IPS). METHODS: In this paper, the main research method is mathematical modeling. As a calculation example, a complex assessment of the technical condition of current transformers and the creation of a predictive model are researched. A predictive model is created based on the existing condition monitoring system (sensors) and methods (digital twin and machine learning). These methods formalize expert knowledge, find implicit correlations, and automate the decision-making process. This paper contains a state prediction algorithm with the ability to build a linear regression model, «decision tree» and «random forest». Researched equipment is a group of single-phase current transformers 110 kV type TFND-110M II. RESULTS: A database was created to predict the rate of development of a thermal defect in the internal insulation of a group of three single-phase current transformers. The algorithm for predicting the technical condition in the form of a linear regression model, a «decision tree» and a «random forest» showed that the best accuracy indicators were obtained from the linear regression algorithm. CONCLUSION: The results of forecasting the technical condition of transformer equipment obtained in this article indicate that in existing systems there is no sensitivity to various scheme-mode parameters of the IPS. The data obtained because of the simulation helps to give the right recommendations to prevent the development of a defect and extend the life of the equipment.