It is difficult to evaluate the mean time between failures (MTBF) of aero-engine with traditional mathematical mean time method because of its small samples. To solve this problem

the historical test information was regarded as prior information

and prior distribution was determined by the moment equivalent method

then field test data were combined with prior distribution through Bayesian theorem

lastly a Bayesian method to evaluate MTBF of aero-engine was presented. The information used in the new method to evaluate MTBF was expanded. An example showed the MTBF evaluation result with the new method was 302.68h

18.7% higher than that with the traditional mathematical mean time method. The engineering experts believed that the new result is closer to the real value than that by the traditional method.It is feasible to use the new method to evaluate MTBF of aero-engine.