Communication is very important for people like Amyotrophic Lateral Sclerosis (ALS) patients who are not able to speak and write, but not with eye movements. At present, there are already several techniques that can read eye movements using the method based on video that has a high price that using signal calculations with input of images called image processing. This signal requires a camera that supports images with a high degree of accuracy and consideration factors that cause noises. One of these factors is lighting so that this method will require a very large cost. Therefore, an alternative is needed that can make eye movements using bio-signals in the body that issupported by movements using Electrooculography (EOG) that supports the direction of eye movements. This research discusses the signal normalization use Differentiation Method. The system can classify of eye movement 8 directions sen ton 8 LEDs. Based on the results of accuracy for eye movement direction is about 100% and based on the results of average accuracy the whole system by taking 5 samples of people is about 79.25%.