Aiming at the problem that the fatigue damage in electronic devices exhibits high uncertainty under random vibration loads， this paper proposes a probabilistic fatigue model based on frequency-time transformation and rainflow counting techniques to predict the fatigue damage distribution of an avionics device under specific random vibration loading conditions. The proposed method is capable of quickly providing the fatigue damage probability distribution， as well as the mean fatigue life which agrees well with the experimental observations. It has the advantages of high efficiency and strong engineering adaptability for random vibration fatigue damage predictions， especially under the circumstances of complicated power spectral density （PSD） loading conditions.