The severe asymmetric wing damage poses a significant threat to the safety of the flight. To improve the accuracy of damaged aircraft diagnoses, this paper designs a fault diagnosis system based on the Unscented Kalman Filter for wing-damaged aircraft. First, a variable center-of-gravity wing-damaged aircraft model is established by using the non-center of mass method. Then, the Unscented Kalman Filters are designed for the aircraft under the normal and wing damage conditions respective, and the probability of faults is calculated through the Bayesian posterior criterion. Finally, the effectiveness of designed fault diagnosis system is demonstrated, whose results show that it has high accuracy and real-time performance and can achieve satisfied fault diagnosis effects.