Abstract: In recent years, email has turned out to be among the pervasive and cost-effective tools of communication. In the meantime, spam emails have reduced their popularity and become offensive to all individuals and users applying this capability. Email filtering is the first solution to cope with this challenge. This is developed as a special type of text classification. A variety of methods including data mining and attribute selection techniques are considered for solving this problem. Incorporating these techniques will lead to the extraction of critical attributes. In this article, we use two hybrid methods: data mining and attribute optimization algorithm together with decision tree algorithms. Results indicate that proposed method can be highly effective in email’s attribute optimization and classification performance increases.