To strengthen electoral integrity and transparency, the global electoral landscape is shifting towards the adoption of biometric voter registration (BVR) systems to curb duplicate registrations and voter impersonation. While effective in improving internal electoral security, their full potential is realized when seamlessly integrated with continuously updated national civil registry systems. Global experiences reveal varied integration models, successes, and persistent gaps. In Kenya, BVR system adoption marked a major step in electoral modernization. Although the current BVR system demonstrates advanced capabilities in biometric data capture and internal deduplication, it suffers from four key problems: it can register illegal (unverified) voters, retain deceased voters, capture mismatched details, and create duplicate entries. These problems stem from the absence of seamless, real-time integration with foundational civil registries that would enable the system to verify citizenship at the point of registration, systematically flag deceased voters post-registration, and instantly combat duplicate entries. This study developed an improved model of the BVR system whose core operation is integrated in real-time with the national civil registry systems to facilitate maintenance of accurate register of voters. The study applied Design Science Research methodology to formulate and evaluate the improved model architecture. A prototype was developed and tested with simulated data, demonstrating improved accuracy and efficiency in voter register maintenance. The findings offer both a practical solution for election bodies and a framework for designing robust, data-integrated systems in electoral contexts.