The resistivity method can be used for the detection of near ground surface, and can also be employed to monitor dynamic underground targets. For the inversion of monitoring data, the single inversion of different data sets has certain defects. For this reason, on the basis of the traditional resistivity inversion algorithm, the time-lapse resistivity inversion formula was derived and the time-lapse inversion program was realized; for the purpose of demonstrating the superiority of the time-lapse inversion algorithm for imaging of dynamic underground targets, a set of multiple forward models was established, and the simulation data were used for single inversion and time-lapse inversion. The results show that, although both algorithms can draw the dynamic underground targets, the time-lapse inversion algorithm can eliminate random errors contained in different observation data sets and reduce the occurrence of inversion artifacts.