Abstract: An efficient structural approach for recognizing Hindi digits drawn by the mouse is proposed. Our Mouse Gesture Hindi Digit (MGHD) system is designed and tested successfully. Our system deals with representation of shape based on a new boundary Freeman Chain Code (FCC) with eight connectivity and then use templates to recognize the Hindi digit. Freeman chain code techniques are widely used to represent an object because they preserve information such as detecting corners, straight lines. The FCC algorithm has been used to produce vector chain code that represented a thinned binary image of the Hindi digit object. In this paper FCC has been modified to extract the boundary of the shape and specify an area of the object were it has been drawn in an image as first check, then matching the result of the recognizer with the result of matching templates as a second check for improving accuracy of the recognition. The proposed method is tested on a sample of 1350 digits written by 27 different writers selected from different ages, genders and jobs, each one wrote 10 digits 5 times. An experimental result shows high accuracy of about 89.5% on the sample test. Experiments showed that this approach is flexible and can achieve high recognition rate for the shapes of the digits represented in this study.