Currently, E-mail is one of the most important methods of communication. However, the increasing of spam e-mails causes traffic congestion, decreasing productivity, phishing, which has become a serious problem for our society. And the number of spam e-mail is increasing every year. Therefore, spam e-mail filtering is an important, meaningful and challenging topic. The aim of this research is to find an effective solution to filter possible spam e-mails. And as we know, in recent days, there are many techniques that spammers use to avoid spam-detection such as obfuscation techniques. In this case, the following proposed approach uses email content only to build keyword corpus, together with some text processing to handle obfuscation technique. The algorithm was evaluated using the CSDMC2010 SPAM corpus dataset that contained 4327 emails in the training dataset and 4292 emails in the testing dataset. The experimental results show that the proposed algorithm has 92.8% accuracy.