Clustering belongs to the set of mathematical problems which aim at classification of data or objects into related sets or classes. Many different pattern clustering approaches based on the pattern membership model could be used to classify objects within various classes. Different models of Crisp, Hierarchical, Overlapping and Fuzzy clustering algorithms have been developed which serve different purposes. The main deficiency that most of the algorithms face is that the number of clusters for reaching the optimal arrangement is not automatically calculated and needs user intervention. In this paper we propose a fuzzy clustering technique (FACT) which determines the number of appropriate clusters based on the pattern essence. Different experiments for algorithm evaluation were performed which show a much better performance compared with the typical widely used K-means clustering algorithm