Handgrip Dynamometer is a measuring device that is often used in the health sector to determine the level of strength or hand grip resistance. This tool has two main components, which are a handheld stick that is used as a handheld media and a screen in the form of an LED panel to display results of information or the value of the grip strength. However, this tool generally is not able to be integrated by computers or other storage devices. Also, the test results cannot be stored or used in collecting data on a large scale. Based on this problem, the research is intended to perform automatic character recognition on the LED screen so that the displayed value can be computerized. The tool used in this design is MIE Pinch or the Grip Digital Analyzer. In short, the LED panel image is taken ten times for each test (one frame per second) using the camera, where this image will be used as the input image. Then the pre-processing process is carried out to fill the gaps between segments in the characters using morphological operations to become one intact area. After that, the Connected Component Labeling (CCL) process is carried out to segment each number on the panel. Finally, the K-Nearest Neighbor (KNN) method is used to introduce each character using classification techniques with previously trained image data. The test results show that the accuracy rate of LED recognition in the tested scenario has a success rate percentage of 84.48% based on 290 valid images from 300 images taken.