Due to the lack of one universally applicable and commonly used reference method, sample preparation in isoflavone (IF) analysis has been performed by many different methods which renders comparison and quality assessment of published IF contents in foodstuffs difficult. In the present work, the impact of different experimental parameters on the IF concentrations determined in soybeans, tofu, soy drink and textured vegetable protein by different extraction and hydrolysis methods was assessed and IF contents obtained by optimized orthogonal methods were compared. Chromatographic analysis was performed by HPLC-UV-ESI-MS. If possible sources of error - which are also pointed out in this work - are avoided, IF contents obtained by extraction, acid-, base- and enzymatic hydrolysis are similar. However, these sample preparation methods differ in the amount of time, standard compounds and instruments required, ruggedness, and in their applicability to analysis of complex composite samples containing soy as minor ingredient. Enzymatic hydrolysis with Helix pomatia juice after extraction by sonication with first 50, then 80% aqueous acetonitrile in the presence of zinc sulfate heptahydrate and after adjustment to </=10% organic solvent turned out to be the method of choice if only aglucone equivalent contents are required. The advantages of this method are short chromatographic run times, smallest danger of coelution, lowest achievable limits of quantitation and therefore best suitability for work-up of complex composite samples and that only aglucone standards are needed for quantitation.