Relational database design using the concept of functional dependencies (FDs) is addressed. The classical synthesis approach processes a given set of FDs to produce one minimal cover, which is then used to develop a relational schema. However, a given set of FDs can have more than one minimal cover, and different minimal covers can give rise to different relational schemata. An enhancement is proposed to the traditional synthesis algorithm that aids in efficiently determining all minimal covers for a given set of FDs. The algorithm has been implemented using Turbo Pascal on an IBM PC AT. The performance of this algorithm is compared with that of the traditional synthesis algorithm.