The mainlobe interference，especially the self-defense mainlobe interference，seriously affects the radar performance since the signals of the mainlobe and the radar echo generated by a real target are within the same beamwidth.Therefore，suppressing mainlobe interference is one of the hot spots and difficulties in the current radar field.To address this issue，a radar mainlobe interference suppression approach based on signal reconstruction is proposed.Firstly，the pulse signals of mainlobe interference are eliminated.Then，based on the modified iterative method with adaptive thresholding （MIMAT），the interfered pulse signals are reconstructed along the slow time dimension.The simulation results indicate that the proposed approach not only could effectively suppress the mainlobe intermittent interference，but also has high robustness to the amount and period of the interference pulses，which has a promising engineering application prospect.