The papershows the classificationof uncertaintiesin solving problemsof forecasting.Uncertaintiesin solvingpredictionmanifested in the formof probability distributionsin the form ofsubjective probabilitiesinthe form ofinterval uncertainty. Statistical uncertainties arise in the case of measurement error, stochastic external disturbances. To account for this uncertainty type digital filters are used, in particular a Kalman filter. In the absence of statistical observations uncertainty is determined by the forward-looking estimates, averages and EM algorithm. Structural uncertainty is always evaluated using the data. The uncertainty determined about and chooses the best of statistical quality criteria. In case of uncertainty of the amplitude type used fuzzy sets. The algorithm of information processing in the uncertainties of probabilistic nature.