Compressor disk is a main part of aeroengine, and is one critical part of aeroengine. The generally fault reasons of disk are low cycle fatigue, oscillation, flexuosity and wriggle etc. Along with the increasing of push weight rate of modern aeroengine, the stress level of it improved greatly, so low cycle fatigue gradually become the primary fault mode. The location of fatigue is the bottom of mortise where the stress concentration is high. The stress state of the mortise bottom is multiaxial. Traditional fatigue analysis methods are uniaxial. The two methods are apparently different. On the other hand, compressor disk is a system which comprises 24 mortises. Any of the mortises failed will lead the fault of disk; all appearance the disk can be thought as a series system which comprises 24 mortises. But if we assume that the mortises are independent and we compute the reliability of the disk base on the series system reliability model, the calculated result would less than the true value. Why the result is less than the true value? Because the mortises of the disk are not independent each other, namely they are dependent.In this paper, the method that theory combined with experiments is used. Researching the disk failure mechanism and reliability analysis method and the building of disk reliability model, the major work is as follows:(1) Refered to a lot of literatures, studied the disk reliability analysis methods in existence, done some strain control fatigue test of disk material and dealled with the test data.(2) Based on FEA and standard fatigue sample test, using the FEA method, deeply analyzed the disk mortises' contact stress with blade rabbets. With the particularly researching on the stress state of disk mortises, the critical plane method is introduced to evaluate the disk fatigue life. It can sufficient consider the influence of multiaxial state to the fatigue life and has higher precision.(3) Analyzed the sensitivity of the effecting factors of compressor disk. Gain the sensitivity curves of the disk fatigue life to correlated parameters. We can determine those parameters which have more influence to the disk fatigue life through comparing the sensitivities. This can be a base of disks of reliability design and manufacture.(4) Through Monte - Carlo digital simulation, a lot of simulation tests had been done. Using the probability weighted moment method and maximum likelood method, the test