Endocrine disruptors are now of scientific and public concern, because there is increasing evidence of their adverse effects on the health of an intact organism or its progeny and on changes in endocrine function. Although numerous substances have been identified as such chemicals, a huge number of chemicals remain to be tested for their endocrine disrupting capabilities. Because of the time and costs required for animal tests, some theoretical or computer-based method for screening this large number of chemicals is needed to reduce the numbers requiring animal testing. Improved quantitative structure activity relationship (QSAR) models were used for screening in combination with other approaches. New receptor-ligand docking simulations were being tested. There was good correlation between experimental and theoretical binding affinities. A database complex system being developed, which enables one to trace cellular signals triggered by the interaction of receptors with xenobiotic chemicals. Perspectives of computer-based screening methods are discussed.