Abstract: For problems of complex modeling, difficult invalid data collection and low prediction precision of existing residual life prediction methods of mine-used cable, a residual life prediction method of mine-used cable was proposed which was based on gray prediction and multi-scale quantum harmonic oscillator algorithm(MQHOA). Dielectric dissipation factors and aging time under different temperature are collected firstly in accelerated hygrothermal aging test. Then residual life prediction model of mine-used cable is established by use of gray prediction GM(1,1) model and the collected data, which takes dielectric dissipation factor of 5% as ending criterion of cable life. Finally, MQHOA is used to optimize parameter of the GM(1,1) model to improve prediction precision. The test result shows that the method has short test time and higher correctness.