Software Architecture Recovery is one of the finest parts of reverse engineering. Several different techniques have been adopted in vast literature to recover Software Architecture. One of the techniques is clustering, which extracts the similar components from the software. In general the component characteristics and its state are not clear. For that, Architecture recovered from ordinary clustering will not be appropriate in it. In this paper we used a fuzzy clustering technique to make the Software Architecture recovery to be more efficient and accurate. Our experimental results have shown that Architecture recovers from fuzzy clustering is better than ordinary clustering.