Abstract: To study the influence of milling parameters on milling temperature in milling GH4169,an orthogonal test between milling parameters and milling temperature was designed. Based on the data obtained from the orthogonal test, the prediction models of GH4169 milling temperature were established by using radial basis function neural network,back propagation neural network and generalized regression neural network algorithms respectively, and the models were trained, tested and verified.In addition,the influence laws of milling parameters on milling temperature was analyzed by the range analysis method. The results showed that the average prediction error of GH4169 milling temperature prediction models based on RBF neural network、GRNN neural network、BP neural network were 3.27%、4.24%、5.05%, and the the average prediction accuracy were 96.73%、95.76%、94.95%. The model based on radial basis function neural network had the highest prediction accuracy, followed by back-propagation neural network model and generalized regression neural network model.The milling temperature of GH4169 increased with the increase of milling speed, feed per tooth and milling depth; The most important factor affecting milling temperature was milling speed, followed by feed rate per tooth and milling depth.