Based on the normalization of measurement data equations of the inverse scattering problem, a new regularization matrix is proposed. It can eliminate the unfavorable effects caused by difference of distances between field-point or source-point and target region, reduce the loss of useful information in regularization procedure, and decrease condition numbers of the ill-posed problems. Inversion for conductivity distribution of two-dimensional axisymmetric inhomogeneous media is carried out by combing this new regularization method with distorted Born iterative method. Simulation results show that compared with the conventional method, the new regularization method is of better stability, quicker convergence, higher accuracy of inversion and higher resolution. Based on the normalization of measurement data equations of the inverse scattering problem, a new regularization matrix is ​​proposed. It can eliminate the unfavorable effects caused by difference of distances between field-point or source-point and target region, reduce the loss of useful information in regularization procedure, and decrease condition numbers of the ill-posed problems. Inversion for conductivity distribution of two-dimensional axisymmetric inhomogeneous media is carried out by combing this new regularization method with distorted Born iterative method. Simulation results show that compared with the conventional method , the new regularization method is of better stability, quicker convergence, higher accuracy of inversion and higher resolution.