A multimodal approach to biometric authentication leverages multiple physiological and behavioural characteristics to verify an individual’s identity. Combining different modalities, such as facial recognition, fingerprint scanning, and voice recognition, can enhance the security and accuracy of authentication systems. Multimodal biometric systems are more resilient to spoofing and provide a higher assurance level than single-modal systems. They are increasingly adopted in various applications, from secure access control in corporate environments to personal device authentication, offering a seamless and robust user experience while safeguarding sensitive information. Our research focuses on developing an original method for multimodal biometric authentication, combining the recognition of keystroke dynamics and facial biometrics. In both cases, neural networks were used, and the necessary data was obtained from public databases. The VGG-Face2 database was used for facial recognition, while the “CMU Keystroke Dynamics Benchmark Dataset” was used for recognizing keystroke dynamics. These modalities were then fused at the match score level. Our developed method for the universal measures AUC and EER reached 0.999882 and 0.004515, respectively. The FAR and FRR measures were 0.004069 and 0.004412, respectively. These results mostly exceed previous achievements described in the literature. The developed solution is noninvasive, enabling its implementation in office environments as an additional layer of protection against unauthorized access to digital devices without requiring additional actions from end users.