Keystroke dynamics is a new security enhancement approach. It relies on identifying users with respect to their distinct typing patterns. However, while many are still familiar with those traditional password-based systems, they are becoming more vulnerable to different types of attacks, pointing out how these systems are getting out of date. Behavioral biometrics used by keystroke dynamics includes dwell time, flight time, typing speed and some other factors. The reliability of keystroke dynamics for user authentication is evaluated in this study through the use of machine learning algorithms. Using Random Forest algorithm, we achieved excellent accuracy of differentiating users based on their typing behavior, which may be useful for practical applications. Keystroke dynamics presents a low cost, user friendly and innovative option to improve security of accounts. This can bring about a true paradigm shift in authentication systems by spanning multiple devices while enhancing security and additionally resisting attacks.