An artificial immune algorithm is applied to blind source separation. By elaborated on blind source separation procedure, the blind source separation based on immune optimization algorithm is put forward, called the AIS-ICA algorithm. Simulation experiments of mixing and separation for four specific signals are carried out. The experimental results show that the convergence speed and the separation precision are high, and it has good stability. The new algorithm is applied to gearbox vibration signals for blind source separation and fault diagnosis, fault information by vibration signals is enhanced. Results show that the algorithm used to separate vibration signals of gearbox can enhance fault information and reduce difficulty of gearbox fault diagnosis.