The increasing adoption of digital electoral systems has introduced significant improvements in accessibility, efficiency, and transparency; however, it has also amplified concerns regarding security, identity fraud, and system integrity. This paper proposes a secure Android-based electronic voting (e-voting) architecture that integrates facial recognition as a biometric authentication mechanism to enhance voter verification and prevent electoral fraud. The proposed system combines mobile computing frameworks, biometric recognition techniques, and secure communication protocols to establish a reliable and tamper-resistant voting environment. Existing literature highlights persistent vulnerabilities in mobile voting systems, including identity spoofing, data interception, and tabulation manipulation (Gupta & Sharma, 2017; Smith & Doe, 2018). To address these issues, the proposed architecture leverages facial recognition algorithms alongside encrypted authentication workflows to ensure only legitimate voters can participate in the electoral process. The study further incorporates insights from biometric authentication frameworks and mobile system design principles to construct a scalable and secure solution (Jain et al., 2011; Li & Jain, 2005). Additionally, the paper emphasizes secure transmission, real-time synchronization, and fraud detection mechanisms to strengthen system resilience. The findings suggest that integrating facial recognition within Android-based voting platforms significantly reduces impersonation risks and improves authentication accuracy, thereby enhancing overall electoral trustworthiness.