Immune quantum optimization theory is applied to independent component analysis. By elaborated on blind source separation procedure based on quantum optimization algorithm is put forward,that is QICA algorithm. Simulation experiments of mixing and separation for three specific signals are carried out. The experimental results show that it has good separation effect. The new algorithm is used to gearbox vibration signals for blind source separation and fault diagnosis. Results show that the algorithm used to separate vibration signals of gearbox can greatly enhance fault information, reducing difficulty of gearbox fault diagnosis.