In order to solve electromagnetic inverse scattering problem, a real-time inverse scattering method is proposed. This technique converts the inverse scattering problem into a regressed one using support vector machine (SVM). Electromagnetic inverse scattering method based on the SVM deals with nonlinearity and ill-posedness inherent in the inverse scattering problem successfully. The feasibility and the validity are tested by making use of simulating through-wall problem, and the results demonstrate that this approach can detect and position the targets behind the wall, no matter whether there exists noise or not. In the through-wall scenario, the influence of the number of sampling positions of receiving antenna on the predicted results is discussed using the predicted model of SVM. In the end, the predicted errors are analyzed and investigated in the multiple source scenario. The results show that this kind of setting is helpful for target identification in through-wall problem.