The present paper describes the design and implementation of an Isolated Word Recognition System using a Z-80 microprocessor. Isolated words spoken by a person are analyzed by a speech pre-processor. It pereforms a short-term Fourier analysis of the signal with the help of a bank of band pass filters. The processed signal is digitized and stored in a RAM for further processing. Templates of the time normalized data are compressed and used as reference or test samples. Pattern matching is based on 'distance measuring' technique for data from each channel. The recognised word is displayed visually and it can also be used to actuate control circuits. The system is speaker dependent. Its accuracy is of the order of 96% when a vocabulary of 30 words is used.