The data of damage inspection and maintenance for aero-engine turbine blades is gathered to form the failure probability distribution function which can be used for inverse analysis by the Paris equation. Then the relationship of detectable crack length and initial inspection time under the specified threshold is gained. After that the degradation state of blades at each inspection moment is determined by simulating the crack growth process. Finally, the probability of failure of blades is obtained via statistical analysis for simulation results. A case is shown that the initial inspection should not be carried out too early because the crack would be too small to be easily detected and the blades would have failed at the subsequent longer repeated inspection moments. Similarly, if the initial inspection is too late to function, the damage size may be close to,or even exceed to, the allowable critical threshold and the probability of failure would be greatly increased at the initial inspection moment. On the condition of 10^-5 allowable probability of failure, the optimal inspection number of times, initial inspection time and repeated inspection interval are both 12 times, and their scheduled operation cycles of turbine blades are 1 371 cycles and 307 cycles, respectively.The method and results can be acted as proofs and references on risk judgment and maintenance decision of aero-engine turbine blades for airline accendant and engineers.