Accurate analysis and prediction of the transformer's thermal condition depending on the operating mode, for example, in cold winters with a shortage of electricity, allows for effective planning of regular maintenance. In the course of the work, mathematical models were created to analyze the thermal state of the transformer, in particular, models for finding the temperature of the upper layers of oil and the highest temperature on the transformer winding. The data from these mathematical models were verified by comparing them with an already identified analog model. It was determined that the difference between the results is no more than 7 %. It has been established that the thermal state of the transformer is influenced by the ambient temperature much more than by the load. This is due to the fact that without cases of overload and emergencies, the load on the transformer, depending on the season, does not change significantly. It has been determined that the highest time utilization and the highest temperature on the high and low voltage windings are observed in August, which coincides with the peak ambient temperature. The lowest temperature on the windings and the lowest life utilization of the transformer are observed in January, which also correlates with the lowest ambient temperature. It is determined that under such operating conditions, given that the nominal service life of the transformer is 20 years, the actual service life will be approximately 90 years. It was also found that the reduction in service life in winter is 5 times less than in summer. This allows us to predict a reduction in maintenance needs during the cold months and more intensive maintenance in the summer. In addition, such models can predict potential problems and emergencies, which can significantly reduce the risk of unforeseen outages and increase the reliability of power supply. Regular monitoring and analysis of the thermal condition of the transformer makes it possible to respond quickly to changes in operating conditions and make timely maintenance decisions, which helps to optimize costs and increase the efficiency of power grids.