The research paper is focused on the assessment of the usefulness of adaptive methods in forecasting demographic variables. The goal of the paper is to conduct the retro and prospective analysis of selected demographic values in the sphere of changes in time, and also to indicate an efficient method for the forecasting of the studied values in subsequent periods. The time series for Poland for the period between 2000 and 2013 are the basis for the development of the forecast. Mean squared errors of ex post forecasts are used as forecast quality measures. The results of the study show that among the applied methods of forecasting, the method of creeping trend with harmonic weights is the most suitable as it gives the smallest forecast errors.