At the present time, the high enlargement of spam e-mail is not easy to preclude and the spam is vastly stretch. These spam e-mails cause various problems to the Internet users, for instance full mailbox, and wasting time. Consequently, prodigious methods have been proposed but most of them have some limitations in mapping features and the processing times. Therefore, this paper proposed a novel method that can detect a set of image spam e- mail called a Partial Image Spam Inspector (PIMSI). The method focuses on spam that consists of both texts and images over the e-mail system. The significant feature of this technique is that it will sensitively protect the distribution of all messages which have a partial similarity of spam e- mail. As a result, users will have a green mailbox as wishes.