Through analyzing the characteristics of the unmanned aerial vehicle (UA) data recorder parameters and the reasons of the outliers, a method using singular value decomposition and wavelet analysis is put forward with the existence of noise to detect outliers with the combination of comprehensive, namely first use singular value threshold de-noising algorithm of signal to noise reduction processing for flight data. After processing with wavelet analysis method for noise reduction of signal detecting outliers, the resulting signal sequence can accurately extract the unmanned aerial vehicle (UA) recorder by the received signal. The simulation results show that the method can not only reduce the noise but also accurately detect and remove outliers existing in the received signal.