Download PDF Abstract Currently, global energy consumption is growing. Traditional energy sources are becoming more efficient, but the growth of the world's population leads to a general increase in energy consumption. Thus, according to the forecasts of the International Energy Agency, the forecast for energy consumption for 2030 will be 33.4 trillion kW * h, and by 2050 it will increase to 41.3 trillion kW * h. In this regard, to ensure the growth of global demands, the energy sector needs fundamental changes, namely, decentralization of generation, introduction of smart grids (Smart Grid), and the use of alternative energy sources (solar energy and wind energy). Only in this case will it be possible to radically reduce the cost of electricity. However, the use of alternative energy sources within the framework of the wholesale electricity and capacity market currently operating on the territory of the Russian Federation is impossible without the use of such short-term day-ahead forecast models. This article analyzed the existing methods of short-term forecasting, which are used to build forecasts of the generation of electric energy in solar power plants. And also, their classification was worked out. To date, there is already a fairly large number of prognostic models built within the framework of each of the selected methods for short-term forecasting, and all of them differ in their features. Therefore, to highlight the most promising method for short-term forecasting for further use and development, an analysis of some of the existing prognostic models was carried out. During the study, the accuracy of forecasting for each of the short-term forecasting methods was evaluated and a conclusion was drawn on the prospects for the use and further development of a hybrid statistical-adaptive method. Keywords Data mining, forecast, electricity, generation, solar power plants, renewable energy sources.