Machine learning has found widespread application across various domains, offering substantial potential for practical solutions. In particular, computer vision tasks such as facial recognition have achieved promising outcomes. This study proposes a student attendance system that integrates a machine learning-based facial identification algorithm with a real-time database to automatically capture, verify, and record students’ presence. The approach employs the HOG (Histogram of Oriented Gradients) method to locate faces within captured images, while Euclidean distance is used to recognize individual students. Facial recognition is performed by extracting key facial features and comparing them with encrypted reference images stored in the database. Each student’s face is captured via webcam, and the system confirms their identity to mark attendance automatically. Experimental results indicate that using HOG for face detection is increasingly efficient and faster compared to Convolutional Neural Networks (CNN).