Abstract: In order to classify audio signals effectively,a new signal feature extraction and classification algorithm is proposed based on SVM(Support Vector Machine) and fuzzy integral in wavelet transform sub-spaces. Firstly,the signal is pre-emphasized and windowed.Then,signals are decomposed into different subspaces using wavelet transform,and features of each subspace are extracted.Thirdly,standardization and dimension reduction are performed to classify signals in each signal subspace.In the end,classification results of each subspace are fused by means of fuzzy integral to get the final class.Experiments show that the proposed algorithm has a faster speed,higher accuracy.