Abstract: The search and localization of unknown radiation sources is an important research topic in the fields of nuclear security inspection and nuclear emergency response. In order to improve the efficiency of source detection and adapt to multi-source environments, a particle filtering method with the Angle of Arrival (AOA) detection is proposed. Firstly, a hardware platform supporting autonomous positioning and AOA sensing is constructed, incorporating detection position and angle information into the search process; secondly, by incorporating AOA information into the particle filtering process, the search area can be dynamically concentrated, significantly reducing computational complexity; finally, the use of AOA to adjust the robot's posture enhances the flexibility of autonomous search path planning. Experiments have shown that this method functions correctly and effectively, and verification with radiation source detection confirms its suitability for searching multiple sources.