With the significant growth of data breaches, companies must think outside to find reliable ways for authentication and identity verification. This paper offers a solution which is to adapt keystroke dynamics authentication (KDA) as a method of two-factor authentication combined with the password instead of using the phone number or e-mail to change a forgotten password or recover a stolen account. The core methodology investigated in this study is the use of Support Vector Data Description (SVDD), a machine learning technique that employs a Gaussian kernel classifier. SVDD is especially useful for anomaly detection, allowing the system to discern between an authorized user’s typing pattern and that of an impostor. This method enhances authentication accuracy while minimizing false acceptance and rejection rates. By combining KDA and SVDD, this study aims to demonstrate that keystroke dynamics can be an efficient biometric authentication component, providing a non-intrusive, continuous, and cost-effective solution. The findings of this study contribute to the emerging subject of behavioral biometrics, paving the way for more secure, user-friendly authentication systems in various applications.