Vulnerability mining has evolved alongside the Internet, initially focusing on exploiting executable vulnerabilities like buffer overflows and heap overflows. As the Web became central to various sectors, it became a prime target for network attacks. Current vulnerability mining focuses on Fuzzing, source code auditing, and rule matching. Automated vulnerability mining, especially with web crawlers and big data, has advanced quickly, but classifying the accumulated data remains a challenge. This paper proposes a machine learning-based Web vulnerability mining system that integrates automated crawling, data collection, feature extraction, classification, and cloud computing technologies.