This paper presents a methodology for enhancing the signature verification system using a Random Forest classifier to rank the feature by their order of importance scores. We focus on finding the most distinct features which contribute to differentiate genuine and forged signatures. We investigate, with the data of SVC2004 that contains a whole range of time-based and dynamic signature features, the systematic use of bootstrapping sampling and decision tree methodologies to rank each feature with respect to their level of impact toward classification accuracy. Apart from this, our methodology reveals essential feature required for the process of successful verification, and it will provide a base to develop more sophisticated and secure technologies for verification of signatures. The finding of this work has the potential to increase the reliability of signature-based systems of authentication and provide a solution to the problems posed by skilled forgeries. Therefore, the current study makes a significant contribution to the larger area of biometric authentication, yielding meaningful insight in terms of ranking analysis of features with an ultimate aim of tightening security levels within the digital environment for personal and organizational data.