The phase-sensitive optical time domain reflectometer (Φ-OTDR) has been widely used in anti-intrusion, crack health detection of building structure and pipeline detection, but the accuracy and anti-noise still need to be improved. In this paper, the Φ-OTDR intelligent recognition system is proposed. Through intelligent identification by neural network algorithm and the vibration intensity is identification with probability processing. The experimental results show that the proposed system can accurately identify and locate the intrusion behavior, and the positioning accuracy is within 5 m