Abstract : An unrolled deep learning scheme for solving the full-wave nonlinear inverse scattering problems (ISPs) is pro- posed. Inspired by the so-called unrolled method, an iterative neural network structure combining contrast source inversion (CSI) method and residual network (ResNet) is designed. By embedding the CSI iterations into the deep learning model, the domain knowledge is well incorporated into the learning process. Thorough numerical tests are carried out to evaluate the performance, stability, robustness and reliability of the proposed approach. Comparisons with the widely used U-net structure and CSI exhibit the advantage of the proposed model.