ABSTRACT: The complex evaluation of a spring as a potential groundwater source includes beside the regime component description also the future prediction of spring yields. This step should be the final and the most important, but of ten it is the most complicated part of the spring yields assessment. In the paper two major problems of spring yields prediction are discussed - input data quality and forecasting methods selection. Probability methods, methods of analytical expression of the relationship among the spring yields and regime influencing factors and methods of analyticaI expression and extrapolation of the time series were chosen as suitable methods for spring yields forecasting. All the used methods were illustrated by examples.