Abstract : At present, many businesses use the advertising emails to communicate with their customers for various objec- tives. Often advertising emails are sent to unwanted recipients that might be from not up-to-date recipient list or bad senders. Currently, the methods of classifying the sender whether they are good sender or not have still been ineffective. Such events can lead to annoyance to the recipients. The recipients might permanently deny to accept these emails or even worse, the business image of email sender could be damaged. Hence, we propose the use of the centralized user feedback database with the sender reputation system to solve the problems previously mentioned. We use the association analysis concept to receive advertising emails from senders by considering relationship between the recipients and senders. Besides, to classify senders accurately, sending histories and the certificated authorities (CA) or the autonomous system numbers (ASN) are also considered. The proposed method can improve significantly toward the classifying process. For the framework validation, the results show that the detection accuracy rate is increased 30.50%.