Fingerprinting is an identification method used by enterprises to personalize services for their end-users and detect online fraud or by adversaries to launch targeted attacks. Various tools have been proposed to protect online users from undesired identification probes to enhance the privacy and security of the users. However, we have observed that new fingerprinting methods can easily evade the existing protection mechanisms. This paper presents a runtime fingerprinting detection and prevention approach, called FPGuard. FPGuard relies on the analysis of predefined metrics to identify fingerprinting attempts. While FPGuard’s detection capability is evaluated using the top 10,000 Alexa websites, its prevention mechanism is evaluated against four fingerprinting providers. Our evaluation results show that FPGuard can effectively recognize and mitigate fingerprinting-related activities and distinguish normal from abnormal webpages (or fingerprinters).