It is hysteretic to use reliability model which is set up based on failure data to reflect the changes of system reliability condition. With the enhancement of reliability level, the early failure samples should be processed time effectively in aero-engine reliability assessment. An attenuation factor α was proposed to express the attenuation grade of failure intensity. The attenuation function that describes the attenuation process was captured by improving Gompertz reliability growth model. All useful information was utilized by using Bayes data fusion methods for time effectiveness of samples. A new acquisition turbo-fan engine was statistically analyzed by this model, and the time effectiveness of fault samples for typical failure was studied. The result indicates that this model could provide more scientific evaluation of aero-engine-s reliability.