When the usages of electronic mail continue, unsolicited bulk email also continues to grow. These unsolicited bulk emails occupies server storage space and consumes large amount of network bandwidth. To overcome this serious problem, Anti-spam filters become a common component of internet security. Recently, Image spamming is a new kind of method of email spamming in which the text is embedded in image or picture files. Identifying and preventing spam is one of the top challenges in the internet world. Many approaches for identifying image spam have been established in literature. The artificial neural network is an effective classification method for solving feature extraction problems. In this paper we present an experimental system for the classification of image spam by considering statistical image feature histogram and mean value of an block of image. A comparative study of image classification based on color histogram and mean value is presented in this paper. The experimental result shows the performance of the proposed system and it achieves best results with minimum false positive.