At present, one-dimensional regularisation inversion is the main way to extract underground electrical structure information in mainstream Helicopter-borne Transient Electromagnetic (HTEM) detections. The reliability requirements for the initial model of the regularisation inversion method is relatively low, and the inversion could be appropriate for different types of targets by choosing different stable functionals, for example, the minimum gradient support functional is generally considered to be suitable for objects with sharp boundaries. However, when there are several low-resistivity layers within a limited depth region, the existing methods have still difficulty in distinguishing each layer. For this reason, this paper draws on the layer stripping inversion concept in seismic data inversion, and proposes a piecewise locking inversion method based on the existing regularisation inversion. This method combines the advantages of the regularisation inversion and the Marquardt inversion, and avoids the interference of late data to shallow target recognition by data segmentation. Lastly, through the application for the simulated and measured data in a lead-zinc mining area, this paper proves that this method can effectively improve the resolution of the HTEM one-dimensional inversion to the adjacent low-resistivity layers.