Software vulnerability remains a serious challenge to the software engineering domain over the past decade as a result of the recent technological advancement in information systems. The rapid development in software applications and failure on the part of system developers to properly analyze program codes before been released to the market increases the chance for data breaches. It is a known fact that most system failures are as a result of bugs and errors detected in software applications. Although code errors significantly affect software quality, there is no effective method that can be used in eliminating software errors to improve its reliability. Data Mining and its related algorithms are an active area which can successfully be applied in analyzing software vulnerability. However the concept of applying data mining techniques has not been empirically proven as an effective method for obtaining the essential characteristics of software vulnerability. To investigate this effect, we propose a vulnerability mining algorithm to analyze and obtain the essential characteristics of Software vulnerability based data mining techniques. We first extracted and preprocessed the software vulnerabilities using data mining techniques and common vulnerability database. We evaluate the proposed technique using the Common Vulnerability and Exposure (CVE) Database, Common Weakness Enumeration (CWE) Database, National Vulnerability Database (NVD) datasets. Empirical results show that the proposed vulnerability mining algorithm has a remarkable improvement in the vulnerability mining process. The most interesting finding is that, we observed that across all the three projects, recall was around 70% and precision was approximately 60%.