The extensive expansion and use of the Web and the Internet comes at the price of seducing a number of intruders to utilize the Web for destructive purposes. In the scope of Websites and Web pages, spammers try to inject their own content and pages in Web sites and search engine search results to be more visible to users and attract users to their Websites or products. This paper analyses the behaviors of the spammers in the content-based Arabic Web pages, through analyzing the weights of the most ten popular Arabic words used by Arab users in their queries. The results show that the behavior of the spammers in the Arabic Web pages can be unique and distinguished in comparison to other languages. Decision Tree was used to evaluate this behavior and it obtains the degree of accuracy which is equal to 90%.