Optical character recognition (OCR) is a type of document image analysis where scanned digital image that contains either machine printed or handwritten script input into an OCR software engine and translating it into an editable machine readable digital text format. In this paper we designed a novel and robust two stage recognition system for Odia handwritten characters as well as we prepare a standard deviation and zone centroid average distance based feature matrix for more accuracy while training and testing the Neural Network. The OHCR System is based on the algorithm of feed forward BPNN in two stage to perform the optimum feature extraction and recognition. The Odia characters are classified into four groups according to similarity of their shapes and features. The system uses ANN in two stages, having different parameters, the first stage classifies the characters into similar groups and in the second stage individual characters are recognized. © 2013 Springer-Verlag.