The motivation that phraseological homosemes are found in the language in a limited number is explained by the following facts: the identical combinatorial probability of the elements of two or more stable polilexical units (SPUs); the small occurrence of the phraseology into the language, and, consequently the non-accumulation of new semes; the semic structure of SPUs is complex and consistent, and ,therefore, it does not allow doubling. According to the ways of formation of homosemic SPUs, homogeneous and heterogeneous units are distinguished. The homogeneous units were formed following the dissociation of polysemic SPU semantics. The heterogeneous SPUs are those that were formed as a result of the sporadic lexical and grammatical coincidence of the elements of these units. They are divided into total and partial units. The total units coincide in the context of the paradigm of lexical and grammatical forms. The partial units reveal some differences regarding the paradigmatic forms of the component elements when used in the context. According to their structure, homonymous SPUs are similar and dissimilar. As for similar SPUs, both their semantic distributive model and the semantics of two or more expression coincide. As for dissimilar ones, their component does not coincide