The ability of a machine to interpret handwritten characters from sources like paper document, photograph, etc. to some editable computerized form is the first and foremost aim of handwritten character recognition systems. Several challenges are there to construct this kind of systems, including the tasks eg. Digitization, segmentation, normalization, feature extraction, reorganization, and reconstruction. Here, an attempt has been made to recognize individual handwritten characters by the use of artificial neurons with feedback connection. The topology of the network is formed with two layers of artificial neurons, input and output. The neurons of output layer have a feedback connection from their output line.