Biodiversity loss driven by land use and land use change (LU/LUC) is a critical global challenge. To quantify and manage these impacts, the scientific community has developed a range of biodiversity metrics, often integrated into 'biodiversity footprints' using tools like Life Cycle Assessment (LCA) and Environmentally Extended Multiregional Input-Output (EEMRIO). However, these metrics are built upon fundamentally different ecological models—including species-area relationship (SAR) refinements (mcSAR, cSAR, SHR, ECA) and pressure-impact models (Mean Species Abundance, MSA; and Species Threat Abatement and Restoration, STAR). The degree to which these distinct models produce consistent assessments of land use impacts remains largely unknown. This study systematically compares six prominent terrestrial biodiversity metrics (mcSAR, cSAR, SHR, ECA, MSA, and STAR) across multiple ecoregions, land use categories, and species groups. Our methodology involves a standardised data processing approach, including log10 transformation and MaxAbsScaler normalisation, followed by aggregation using the non-zero geometric mean. We then employ a comprehensive statistical framework, including median-based comparisons, ecoregion-level pairwise correlation analyses (Pearson and Spearman), and non-parametric tests (Kruskal–Wallis H and Mann–Whitney U tests), to evaluate the consistency of the metrics. We address two key research questions: (1) To what extent do biodiversity metrics derived from different ecological models provide convergent or divergent assessments of land use impacts? (2) In which terrestrial ecoregions and for which species groups do these biodiversity metrics most significantly align or diverge? By systematically evaluating the strengths and limitations of these underlying ecological models, this work is critical for informing robust policy, conservation planning, and corporate reporting that relies on biodiversity impact metrics.