Abstract: Interference recognition is an important part of the communication electronic defense system. As for the interference recognition problem of the complex electromagnetic environment, this paper studies an interference recognition technology based on the joint multi-dimensional features, extracts a set of features which are less sensitive to interference parameters and noise, and are less complex from the time domain, frequency domain and transform domain. With the joint multi-dimensional features, two kinds of interference identifier structures based on the decision tree and support vector machine (SVM) are proposed, and interference recognition performance simulation is carried out. The simulation results show that these two identifiers have good recognition performance for typical electromagnetic interference, they have similar performance for the aiming interference, the partial band noise interference, the noise frequency modulated interference and the pulse interference; for the comb shaped multi-tone interference and linear frequency modulated interference, the SVM recognizer has better recognition performance than the decision tree recognizer.