In the method according to the invention a phonocardiographic signal, obtained by a passive acoustic sensor (10) from a maternal abdominal wall, is pre-filtered, amplified, digitized, digitally filtered by means of a programmable amplifier and filter unit (20), is stored in an intermediate memory and autocorrelation is performed by means of an autocorrelation unit (30) in a time window of a predetermined size. The signals obtained as a result of autocorrelation are fed to an input of a fuzzy expert system (40), wherein local maximums of the signals, temporal location of the local maximums and variation of the temporal location of the local maximums are determined by the fuzzy expert system (40), and, applying those as input parameters or signals of the fuzzy expert system, those are classified into probability groups - utilizing a fuzzy rule set of a decision logic stored in a knowledge base and biologically expected data - and are evaluated.
