Displacement information of a moving target can be detected using an optical feedback self-mixing interferometry (OFSMI) system. A sensing signal observed from the OFSMI system is called self-mixing signal (SMS). The paper studies the waveform features of the SMSs and proposes an algorithm for reconstructing the displacement of a moving target. The reconstruction accuracy of the algorithm mainly depends on the locating accuracy for those characteristic points on a SMS. A set of rules for identifying those characteristic points are described in the paper. The proposed algorithm is verified by simulation signals firstly, and then applied on extensive SMSs which are obtained from the experimental set-up. The results show that the displacement of the external moving target can be reconstructed under different feedback levels.