Abstract:Aimed at the problems of small sample,low generalization, overfitting and local minimum, a milling force prediction method was established based on Support Vector Machine (SVM). It was used to predict the milling force, and compared the predicted results with the test value and the BP neural network predictive value. The result shows under the circumstance of small training samples, the SVM predictive value gets a good agreement with the test value, and comparing with the BP Neural Network training, SVM gets higher prediction results.