Abstract: Currently,artificial prosthetic technology has become a research hotspot,the identification of gesture surface electromyography signal made a significant contribution to the study of artificial prostheses.This paper proposed a signal processing method based on Ensemble Empirical Mode Decomposition(EEMD),which firstly decomposed the EMG signal into several Intrinsic Mode Functions(IMF) and then extractied the time domain feature,frequency domain feature and Hilbert domain feature from the first three IMF.The extracted features composed a multi domain feature vector.Subsequently,the multi domain feature vector was input to Support Vector Machines(SVM) to classify six types of gesture actions,the mathod reached highest recognition rate,which was 98.7%.Experimental results showed,using EEMD to decompose the original signal to extract multi domain features,was feasible to identify gestures.