Abstract: The fragile ecological environment of the Loess Plateau significantly increases the susceptibility to rainfall erosion, underscoring the importance of accurately assessing rainfall erosivity to optimize ecological management strategies in this region. However, discrepancies in rainfall erosivity estimation models have led to variations in spatiotemporal patterns and driving mechanisms, complicating the understanding of erosion dynamics. To address this, this study utilized high-resolution hourly rainfall data from 281 stations across the Loess Plateau spanning the period from 1990 to 2021. The research aimed to analyze the differences between rainfall erosivity models based on rainfall amount and rainfall intensity, and to explore the spatiotemporal distribution patterns of rainfall erosivity derived from these models. Furthermore, the study investigated the influence of key factors such as seasonal variations, rainfall types, and elevation on the formation and distribution of rainfall erosivity.The results revealed several critical findings: (1) The annual average rainfall erosivity calculated using the rainfall intensity model was 918.36MJ mm hm?2 h?1, with values ranging from 222.81 to 2583.52MJ mm hm?2 h?1. In contrast, the annual average rainfall erosivity derived from the rainfall amount model was 1058.17MJ mm hm?2 h?1, with a range of 271.92 to 2700.71MJ mm hm?2 h?1. (2) While both models yielded comparable spatial and temporal patterns in rainfall erosivity, the model predicated on rainfall intensity exhibited a greater degree of variability, as evidenced by its higher coefficient of variation (CV) of 0.44. This elevated CV underscores the heightened sensitivity of the rainfall intensity model to interannual fluctuations and variations in the characteristics of rainfall events, suggesting that it is more responsive to changes in rainfall patterns over time. (3) The spatial patterns of rainfall erosivity were consistent across different rainfall types and seasons, with the highest values and most extensive distribution observed during the summer months. Additionally, the rainfall intensity model demonstrated greater sensitivity to elevation, as evidenced by a correlation coefficient of -0.58, suggesting that elevation plays a significant role in influencing rainfall erosivity when calculated using this model. In conclusion, this study highlights the importance of integrating the dual effects of rainfall amount and intensity to accurately capture the dynamic characteristics of rainfall erosion processes on the Loess Plateau. By providing a more precise evaluation of rainfall erosivity and its response to regional parameter variations, this research supports efforts to conserve the fragile ecological environment of the Loess Plateau. Furthermore, the findings offer valuable insights for policymakers in formulating effective soil and water conservation strategies, ultimately contributing to sustainable land management and ecological restoration in the region.