A Form-based Optical Word Recognition (OWR) System for printed forms includes functional mechanism for extract word from image and word image recognition. The recognizer is a very important component of the OWR system. Automatic recognition for printed images is a complex task. In this paper we discuss the form image recognition technique and improved field matching techniques implemented in our system. These effective techniques help in preparing the input word image for the Hidden Markov Model (HMM) based recognizer and Partial Matching (PM) algorithms for improving overall recognize matching accuracy. Although these algorithms have been implemented with Liver Cell reports in our OWR.