The received signals from ground-penetrating radar (GPR) contain round-trip echoes, clutters, and complex noise signals. These jamming signals seriously affect the interpretation precision of shallow geological subsurface information. In order to dissolve some useless signals in GPR signals, it is necessary to take appropriate measures to repress interference. Based on the electromagnetic field theory, the propagation characteristics of the transmitted GPR signal are analyzed. On this basis, a matched filtering algorithm based on phase-shifting pursuit is proposed to enhance the received GPR signals. At first, the intrinsic component libraries (ICL) can be generated by changing the phase of the transmitted GPR signal. Then, the correlation analysis between the local information of the received GPR signals extracted by sliding window method and each sample in ICL is studied to extract target echo signals. Experiments based on the GPR imaging demonstrate that the proposed algorithm could enhance the target echo signals to a certain extent. The integrated side lobe ratio of the imaging result of the enhanced GPR signals is 6.33 dB lower than the original ones. The resolution of target imaging can be improved.