Squared-error Pattern Clustering algorithm used in unsupervised pattern recognition and image processing application demands substantial processing time for operation of feature vector matrix. So, this paper propose the fast squared-error Pattern Clustering Processor using the Residue Number System which have been the nature of parallel processing and pipeline. The proposed Squared-error Pattern Clustering Processor illustrate satisfiable error rate for Cluster number which can be divide meaningful region and about 200 times faster than 80287 coprocessor from experiments result of image segmentation. In this result, it is useful to real-time processing application for large data.