Drowsiness causes changes in the behavior of the eyes, mouth, and brain. In a sleepy state, the eyelids close together, the mouth opens slightly, and the brain generates low-frequency signals. These changes allow you to detect drowsiness. In our research, we build a machine learning model to detect eye conditions. We used a public dataset of 4000 images where there are 2000 images of completely closed eyes and 2000 open eyes. All data is trained on Convolutional Neural Networks (CNN), creating a model for testing. In addition, testing was carried out on three more algorithms such as Logistic regression, Random Forest and SVM. The greatest success (97% accuracy) was achieved using the CNN algorithm. Despite the encouraging results, our study revealed certain limitations and challenges. Variations in blink rate and the complex relationships between blink patterns and levels of sleepiness pose complexities that were not fully accounted for in this study.   Understanding these limitations is necessary to improve fatigue detection systems.