Laterolog (DLL) logging, utilizing deep and shallow detection depths, is instrumental in identifying and evaluating complex reservoirs such as coal measures. It effectively captures the resistivity contrast between the virgin zone and the invaded zone, providing crucial data for fluid identification, reservoir quality assessment, and fracture system characterization. However, the DLL response in practice is a complex function of coupled factors including borehole environment and bed thickness. Existing studies have predominantly focused on conventional sandstone or carbonate reservoirs, leaving the response characteristics in low-resistivity reservoirs like coal seams poorly understood, which consequently limits the accuracy of quantitative evaluations in fractured reservoirs. To address this gap, this study established a two-dimensional numerical model for deep and shallow laterolog responses using the COMSOL finite-element platform. We systematically investigated the influence of key parameters—including sonde coefficients, borehole diameter, coal seam thickness, and fracture properties (count, width, and spacing)—on DLL resistivity responses. The results indicate that sonde coefficients are significantly influenced by model dimensions; the shallow laterolog sonde coefficient stabilizes rapidly within the detection range, whereas the deep laterolog sonde coefficient requires a substantially larger model domain to achieve convergence. An increase in borehole diameter causes a simultaneous decrease in both deep and shallow resistivities, characterized by a “positive separation” where the shallow resistivity exhibits a more pronounced drop. A positive correlation exists between coal seam thickness and apparent resistivity; the vertical resolution of the DLL for coal seams is approximately 2.5 m, below which the seam’s response characteristics are readily masked by adjacent formations. Fracture development markedly alters resistivity curve profiles, inducing characteristic depressions. Increasing the number of fractures or their width exacerbates these depressions and lowers the overall resistivity. The shallow laterolog curve often displays a distinct “spike-depression” morphology at fracture zones. Notably, increased fracture spacing facilitates the formation of macroscopic conductive pathways, leading to a further reduction in resistivity. The simulation results demonstrate excellent agreement with field-measured data, validating the high reliability of the model. The findings provide a robust theoretical foundation and methodological reference for refined DLL interpretation, environmental correction, and reservoir quality assessment in coal seams and fractured reservoirs.