Recording student attendance in classes at the National Development University "Veteran" East Java is currently still done manually. This is considered less efficient. With the rapid development of today's technology, attendance systems that utilize biometrics should be created and developed. In this research paper, the researcher presents the proposed design of a facial recognition-based biometric presence system. The research method used is literature study, current system analysis, system requirements analysis, and system design. The learning method used is Convolutional Neural Network (CNN) by utilizing the OpenCV and tensorflow libraries. The implementation of the system will utilize CCTV surveillance cameras that are in the class of the Computer Science Faculty of UPN "Veteran" Jawa Timur.