Abstract: With the acceleration of urbanization, shield construction has been widely applied in underground tunnel engineering. However, due to the complexity and uncertainty of underground geological conditions, shield construction often faces many risks, such as water gushing, mud burst, collapse, and rock burst, which not only affect the excavation efficiency, cause delays in project progress, and result in huge economic losses, but also seriously threaten the lives of construction workers. Therefore, accurately predicting the location of water-bearing structures ahead of the tunnel face over long distances is an urgent exploration challenge to be solved. The DC resistivity method is recognized as a sensitive detection means for water bodies and voids and other adverse geological bodies. The focusing detection method based on the steady-state field is an extension of the DC resistivity measurement technology, and research has been conducted on the prediction of the location of abnormal bodies ahead of the tunnel face. Firstly, considering the frequency dispersion effect of the stratum, forward modeling was carried out for different strata with adverse geological bodies to explore their resistivity response characteristics. When a low-resistivity abnormal body exists in a high-resistivity stratum, the larger the ratio of the stratum resistivity ρ1 to the abnormal body resistivity ρ2, and the smaller the change trend of the percent frequency effect (PFE); when a high-resistivity abnormal body exists in a low-resistivity stratum, the larger the abnormal body resistivity, the greater the change trend of the apparent resistivity, and the greater the change trend of the PFE. Secondly, by using the voltage difference information between the layer with the abnormal body and the homogeneous layer in the early stage of excavation, the voltage difference information ahead of the tunnel face was optimized using the gradient descent method. It was found that the positions of the trough and peak were not affected by the stratum resistivity and the abnormal body resistivity. As the tunnel face approached the abnormal body, the trough and peak phenomena became more obvious, and the distance L between the trough and peak positions and the tunnel face gradually decreased. Finally, based on the relationship between L and the position x of the tunnel face, the location of the abnormal body was predicted. When the abnormal body was far from the tunnel face, the prediction results were poor. When the abnormal body was within 38 m ahead of the tunnel face, the prediction error could be within 3 m, which improves the advance detection distance when there is water-bearing structure in front of the face of the tunnel.