This paper purports a refreshing spam discovery technology for chat system based on rough set theory. Nowadays, spam is very much allied with a huge chunk of data transferred through internet involving all disturbing and unsolicited contents received via different web-services such as chat systems, e-mail, forums and web logs. In this paper, we have reviewed various past research works of filtering SPAM and propose a novel filtering technique for SPAM especially for chat system with the support of classical rough set theory. Simulation results clearly indicate that our proposed method, can achieve higher accuracy in spam detection as compared to the existing strategies. © 2013 IEEE.