Average daily or monthly data of solar radiation are often used in statistical analysis and in development of mathematical models for predicting energy generation. However, this can have a negative impact on the accuracy of forecasting. The authors carried out the statistical analysis of solar radiation based on the data obtained from the Tokmak SPS of the Zaporizhia region, Ukraine during 2016. The data were collected with a 30-minute interval in measurement. The analysis indicates close values of winter and summer insulation indices (direct and scattered), where the difference is approximately double. The mathematical expectation of the power surges is close to zero, and their distribution has signs of symmetry. The results of actual data calculation indicate that values of the elementary power surges of the SPS are a bit higher than the normal distribution calculated. Thus, the actual distribution corresponds to the normal one in the parameter of symmetry, but does not correspond to the parameters of kurtosis. If the time range is doubled from 30 minutes to 1 hour, the range of power changes slightly increases. In addition, the analysis of meteorological data showed that the probability of a prolonged growth or decline in power is not high, i.e. a change in power as a random process can be considered close to the stationary one. The performed analysis estimates the power supply possibilities for the region’s consumers by the solar power plant, and creates a mathematical model for operation of the combined energy system that uses solar energy.