Abstract: The acoustic emission (AE) real-time monitoring for fatigue damages to the horizontal tails and some inaccessible components was carried out during the functional test for a certain plane. In a noise background due to severe shock and vibration, the simulated AE sources and background noise were analyzed to pre-process the acquired data properly so as to enhance greatly the signal-to-noise ratio (SNR). Then, a trend analysis was made for the characteristic parameters of AE signals. The test results showed that in a long-time fatigue test process the background noise may change correspondingly to the change in ambient conditions, and the change in the trend analysis of the characteristic parameters of AE signals will therefore reflect properly the randomness/nonsteadiness in the changing background noise. Illustrated with the half-shaft fatigue damage, the trend of the characteristic parameters of AE signals (in hits) can reflect successfully the history of half-shaft fatigue damage. Thus, the time the initial fatigue damage starts and two symmetrical sources of fatigue damage can both be determined basically by analyzing the waveform of AE signals.