In order to improve the efficiency of the inversion of CSAMT sounding data, this paper suggests that parallel computation should be introduced into the 1-D iterative inversion of CSAMT sounding data for pseudo-2D inversion interpretation. Based on multiprocessor and multi-core workstation and using the programming technology of multithread and multi-process, the authors formed the parallel algorithm and the final program. The parallel algorithm was adopted for inversion interpretation of measured data, and the result shows that the pseudo-2D inversion using parallel algorithm can greatly improve the efficiency of the inversion of CSAMT sounding data.