A quantum genetic algorithm (QGA) with good global optimization ability and fast convergence speed is proposed to solve the parameter selection problems of least squares support vector machine (LSSVM) on wind turbine gearbox fault diagnosis model. The method can convert the LSSVM model parameter selection into optimization. It overcomes the problem that GA is easy to fall into local optimum in the optimization process and it also improves the optimization performance. A series experiments are carried out on the data sets of UCI database. Compared with GA—LSSVM and CV—LSSVM, the classification accuracy is improved. Finally QGA—LSSVM model is applied to the wind turbine gearbox diagnosis and a good result is achieved.