The article proposes a method of estimating forecasting methods (regression–correlation analysis method and statistical dependence equations) based on the analysis of forecast errors using the method of complex statistical coefficients. It is proposed to perform forecasting error by comparing the forecasted and actual values of the indicators. This retrospective approach allows you to establish a better prediction method. Trend calculations are also presented graphically, with minimum, average and maximum forecast values. The application of the method of statistical equations of dependencies for studying changes in dynamics allows to reduce the level of error of predictive calculation due to the fact that such study allows to obtain scientifically sound results in both a small and numerous set of levels of a dynamic series. The reliability of the calculations of the forecast of phenomena and processes on the basis of the method of statistical equations of dependencies is ensured by calculating for the investigated equation the level of stability of the trend. In order to determine the interval forecast values (minimum and maximum forecast values), it is proposed to use the mean linear deviation based on the method of statistical dependence equations.