Abstract: Farmland fragmentation, driven by interactions between natural and anthropogenic factors, manifests as reduced parcel size, shape irregularity, spatial dispersion, and fragmented ownership patterns. In China, it has formed the “small-holder, large-country” agricultural pattern and become a prominent barrier restricting the promotion of modern agriculture. This phenomenon significantly impedes large-scale agricultural modernization and diminishes production efficiency. Although various fragmentation indices have been developed to quantify these patterns, elucidate underlying mechanisms, and inform land consolidation strategies, their development remains challenged by the multi-scale, multi-dimensional, and spatiotemporally heterogeneous nature of fragmentation. Moreover, there is still no universal standardized assessment scheme, which makes it difficult to generate comparable evaluation results across different regions. This study systematically reviewed and evaluated common measurement indices across three spatial scales: geometric indices at the parcel level, landscape and ownership indices at the meso-micro level, and aggregation metrics at the regional level. Literature synthesis and correlation statistical tests were jointly adopted to compare the advantages, defects and applicable scenarios of each indicator. Our review revealed that, despite a relatively mature index system, four critical limitations persist. First, conceptual redundancy and overlapping computational methodologies compromise index objectivity and effectiveness, while the lack of clear quantitative linkages between index values and agricultural productivity or ecological risk leads to ambiguous interpretation. Second, inconsistent value ranges, indicator orientations, and terminology hinder cross-study comparability and knowledge transfer. Third, insufficient attention to scale dependency and regional heterogeneity constrains index applicability and cross-regional comparability. Fourth, current approaches inadequately integrate explicit landscape fragmentation with implicit ownership fragmentation, and fine-scale measurement techniques remain underdeveloped. Landscape-derived indices are relatively accessible yet cannot fully reflect tenure-related fragmentation, while ownership-based indices suffer from limited data availability in practical applications. Most composite comprehensive indices rely on subjective weighting schemes, which further introduce uncertainty into final evaluation outputs. To empirically examine these limitations, we conducted a multi-scale comparative analysis across county, grid, and small-watershed scales in China's black soil region, revealing that scale effects fundamentally alter the meaning and behavior of indices, with some indicators reversing correlations or losing discriminatory power as spatial units change. Notably, weighted average indices such as Shape Index (SI), Mean Shape Index (MSI) and Area-Weighted Mean Shape Index (AWMSI) are highly susceptible to the interference of large dominant parcels and tend to lose discriminatory capacity in small-watershed assessments, which requires cautious application. Based on these findings, we propose six key insights: (1) Integrating fragmentation assessment into cultivated land quality frameworks using scale-specific indices; (2) Establishing region-specific quantitative relationships between geometric indices and productivity-ecological tradeoffs at the parcel scale; (3) Developing coupled landscape-ownership indices to guide targeted consolidation at meso-micro scales; (4) At the regional scale—reflected in aggregation disparities—direct cross-regional comparisons are fundamentally constrained by scale effects and contextual heterogeneity; (5) Theoretically, an optimal fragmentation threshold exists and should underpin the development of region-adaptive, standardized indices; and (6) Empirically, for regional/landscape-scale analysis, the combination of Patch Density (PD), Percentage of Landscape (PLAND), Division Index (DIVISION), and modified Shannon Diversity Index (MSHDI) is recommended, whereas at micro/parcel scales, Patch Density (PD), Fragmentation Shape Index (FS), and Januszewski Index (K) demonstrate superior performance. This review provides theoretical advances for fragmentation measurement methodology and practical guidance for scientific land consolidation, high-standard farmland construction, and sustainable agricultural development.