The dependence space models are applied to representation and investigation of domains with incomplete information. It is shown that various models of these domains can be uniformly. treated within the framework of the theory of dependence spaces. In particular, dependence spaces are constructed for contexts ([Wi1]), information systems ([Pa7]), decision tables. The problems are studied that arise in connection with discovery of knowledge from indiscernibility-type incomplete information. The algorithms are given for realisation of various relevant tasks. Among others, an algorithm for finding reducts of sets of attributes in an information system and an algorithm for reduction of a set of conditions in a decision table are presented.