The construction of intelligent loal mine requires the employment of the intelligent geological exploration technology to detect and predict front area in the process of roadway tunneling. The real time detection system of seismic while excavating is constructed on the basis of on-line seismic monitoring stations. The acquired seismic data is inspired by vibration of the tunneling machine, and is transmitted on optical fiber network to the database server located on the ground. The processing software acquires current seismic data from the database, through the steps of screening, pulsing and migration, imaging the front and side area of roadway head by reflected in-seam wave. In order to verify the performance and effectiveness of this system, we installed the real time detection system in the 5106 air return roadway of the Yushupo Coal Mine, and took a tracking detection test for several months. The detection system collected real-time seismic data, imaging in real time while excavating. With the increasing of the tunneling length, daily detection results show that there is a hidden fault existing in working face 5106. The fault has been verified by later drilling and detection of reflected in-seam wave. That means this real time detection system has the ability to image the area in front and side of the roadway by seismic signal generated from the road-header. Therefore, the detection system realized the continuous tracking and real time monitoring of abnormal structures in front of roadway side without affecting the process of roadway tunneling. It meets the requirement of geological detection ability for an intelligent tunneling system.