Abstract:In order to improve the aeroengine wear prediction accuracy of support vector regression(SVR) prediction model,a improved particle swarm optimization(IPSO) was presented to optimize the structure parameters of SVR and vector dimension of training samples.Nonlinear inertia weight and adaptive detection response mechanism were used in IPSO to overcome the shortcoming that the random initial positions of the traditional particle swarm influences the optimization results,and the accuracy and consistency of the SVR predication results were improved.The wear prediction of an aeroengien was carried out based on the Spectral analysis data.The result shows that,compared with the traditional PSO-SVR and common neural network prediction model,the IPSO-SVR prediction results have higher accuracy and better consistence.