The parallel wire cable is regarded as a series-parallel model with multi-crack competition， and the initial crack size of corroded steel wire is determined by a pitting non-uniformity coefficient. Based on the probability distribution of the parameters of the fatigue crack growth rate model， the fatigue life of the cable under different wire breakage rates is calculated by using the Monte-Carlo simulation method of the whole process of fatigue failure， and the calculated results are verified by the test data. The results show that the wire-breaking rate of the cable conforms to the growth law of the power series with the increase in fatigue times. When the wire-breaking rate of cable reaches about 10%， the wire-breaking rate of cable increases rapidly and loses the ability to resist fatigue load. The more the series steel wire elements are divided in the fiber bundle model， the shorter the fatigue life of the cable， but this trend gradually stabilizes with the increase of the number of steel wire elements. The fatigue life of the corroded cable follows the Weibull distribution， and the pitting effect shortens the crack initiation life of the steel wire， resulting in a significant decrease in both the average fatigue life and variability of the cable with the increase of the degree of corrosion， as well as a rapid deterioration of its anti-fatigue performance. The random load history is transformed into the repeated action of fatigue load spectrum block by the rain flow counting method. Based on the corrosion development law of steel wire， the fatigue life of cables with different corrosion degrees is calculated， and the relationship model between cable corrosion degree and fatigue life degradation under random load is established. According to the P-dc-N curve， the fatigue reliability of corroded cables can be evaluated. It is necessary to accurately grasp the defect or corrosion details of cable steel wire in the application.