We aim to convert the binary cross association clustering as described by Modha, Chakrabarti, Papadimitriou, Faloutsos to a soft clustering method. The binary cross-association (BCA) clustering method is inherently performing hard clustering in that one row is a member of one and only one row cluster and similarly a column can be only part of one column cluster. In our work, we relax this requirement in such a way that each row can be a member of more then one row cluster and each column can be a member of more then one column cluster. In this new method, called soft binary cross association (SBCA) clustering, each row does not belong to exactly one row cluster but belongs to each row cluster with a certain probability or degree of membership and similarly for columns.