The comprehensive index system and the single index system have been widely used to evaluate the ecological vulnerability. However, the systematic quantification of the consistency between the results obtained by these two methods is still lacking and the uncertainties in the spatial patterns and driving forces of the ecological vulnerability remain unclear. In this case study conducted on the Loess Plateau which is one of the most eco-fragile regions in China, we compared the spatial patterns and driving forces of the ecological vulnerability as estimated by a comprehensive index system based on the exposure-sensitivity-adaptation model and a single index system based on the ecosystem net primary production. Results showed an overall consistency in the spatial patterns of the ecological vulnerability results calculated by the two methods. Approximately 58%-63% of the total land area exhibited a moderate or higher ecological vulnerability level, with higher ecological vulnerability observed in the northwest region compared to the southeast region. On a per-pixel basis, although the estimated ecological vulnerability indices by the two methods were highly positively correlated across the space (r=0.80, P < 0.001), nearly half of the lands (45%) showed inconsistent levels of the ecological vulnerability, characterized by one level higher vulnerability by the single index system compared to the comprehensive index system. This inconsistency was particularly noticeable in the southeast, central and west regions with relatively higher vegetation productivity. Our findings revealed that the spatial patterns of the ecological vulnerability estimated by the two methods showed overall consistent correlations with the main driving factors according to the Geodetector model. The climate humidity represented by the annual precipitation or moisture index and the vegetation coverage represented by the annual maximum Normalized Difference Vegetation Index (NDVI) were the two major drivers, with the largest explanatory power by the enhanced interaction between annual precipitation and annual maximum NDVI. On average, these two factors explained 85% of the spatial variations in ecological vulnerability estimated from the two methods. While our findings emphasized the highly consistency of the ecological vulnerability patterns estimated by different methods. They also highlighted some discrepancies, especially in the regions with relatively better vegetation conditions. The single index system might be more cost-effective in evaluating the ecological vulnerability patterns, considering the complexity of calculation methods and the large inter-annual variability of the estimated ecological vulnerability by the comprehensive index system method. This study provided baseline inferences for the selection of calculation methods in follow-up studies on ecological vulnerability evaluation.