Abstract:,。On the basis of analyzing disadvantages of conventional prediction model, a novel prediction model based on modified PSO least square support vector machine was proposed. Based on the new model, the design steps and learning algorithm were given. The practical experimental results show that the construction speed of this modified PSO LS-SVM model is 10 times less than that of the LS-SVM model, while the prediction error is 50%. Moreover, compared with BP model, the prediction accuracy is about 10 times higher than that of the former. The effects of electrostatic spraying parameters on atomization performance of electrostatic spraying nozzle can be predicted with the limited test data. Thus the variation law of atomization performance of electrostatic spraying nozzle following electrostatic spraying parameters can be obtained.