The awareness of terrorism has increased since the event of the September 11 in the U.S. It has focused the spotlight on the absence of information available about terrorism. The fight against terrorism has trigger new effort to find better approaches to intelligence data gathering and analysis. Generally, a major activity in counter-terrorism is the ability to process high-volumes of data using various Artificial Intelligence (AI) techniques to identify potential threats. Recently, there has been a huge increase in the amount of intelligence-surveillance data literature. However, it so difficult to gain access to relevant data when the data are scattered in different departments and exist in many forms such as phone tap scripts, internet usage logs, emails, seized hard-drive, CCTV video, images, etc. In this paper, a survey is presented on the technical challenges in extracting terrorism data and sets forth several significant approaches to these technical challenges. The performance by each reviewed approaches in counter terrorism and other domains is evaluated. The potential of using Link Grammar (LG) in information processing is also explored. © 2009 IEEE.