This paper adopts a method to separate the faulty vibration signal based on the blind sources separation (BSS) method. Firstly, the technique of failure diagnosis of rotor vibration, the main faults of aeroengine, and the mechanisms of the vibration are introduced. Then, the blind source separation technique is chosen and its availability to be implement ed in aeroengine fault diagnosis is analyzed. At last, the vibration character istics of an aeroengine which vibrates seriously is determined. The fast indep endent component analysis(ICA) and java agent development enviroment(JADE) metho ds are applied to separate the source vibration signal from the multisensor si gnals. The results show that the BSS method can separate source vibratio n signals effectively and improve the accuracy of fault diagnosis of aeroengin e.