Representation and manipulation of chemical structures are very important for database systems and expert systems in chemistry. Especially, unique naming3, 6, 7, 9, 11 and substructure matching12 are most important. Usually, a chemical structure is represented as a graph, in which an atom corresponds to a vertex and a chemical bond corresponds to an edge. How-ever, a graph is not sufficient for representing a chemical structure. Stereoisomers must be distinguished. Though they have the same graph structures, their geometric structures are different and they show different properties. How to represent stereoisomers in computer systems has been studied well. Especially, the works of Wipke and Dyott are well known. They developed the stereochemically unique naming algor-ithm13, 14 (stereochemically extended Morgan algorithm, ab-breviated as SEMA) based on the ordered list representation of stereochemical structures by Petrarca et al. 10 In thispaper, another approach to distinguish stereoisomers is presented. A basic technique in computer science “transformation” is employed. A chemical structure is transformed into a usual graph (a structure which does not have stereochemical information). Besides, two structures are transformed into isomorphic graphs, if and only if they are stereochemically isomorphic.