Abstract: Abstract: Splash erosion, which is a main dynamic for the detachment and transport of soil aggregates, is the initial stage of soil water erosion. Soil aggregate is a crucial indicator determining the plough layer soil fertility. Destruction of the plough layer aggregate has a negative effect on the content of soil available water and soil fertility, destroying the soil structure and restricting agricultural development of the Loess Plateau. To determine the effect of raindrop diameters on the breakdown of soil aggregates of plough layer and the characteristic of splashed fragment size distribution in the process of raindrop splash erosion, an artificially simulated raindrop splash experiment was conducted with 4 raindrop diameters (2.67, 3.05, 3.39 and 3.79 mm) using a self-designed simulated device that had a circular room with a 10-cm diameter for placing soil samples. The soil sample was collected from the soil plough layer (0-20 cm) with a steel ring (10 cm in diameter). The sampling site was located in a traditional agricultural planting region in Yangling of Shanxi Province (108°03′29″E, 34°18′24″N). Each raindrop diameter splash test was replicated 3 times. For each raindrop diameter, 5000 raindrops were dripped from a generator device on sample, and splashed aggregate fragments were collected every 1000 raindrops. The splashed fragments were sieved with aperture (2, 1, 0.5, 0.25, 0.106, 0.053 mm) using an aggregate analyzer (HR-TTF-100). All aggregate fragments were oven dried for 24 h at 105°C and weighed. The results showed that the splash amount among the collected order was not significantly different for raindrop diameter 2.67 and 3.05 mm, whereas, the splash amount of the II-V order was significantly lower than that of the I order for raindrop diameter 3.39 and 3.79 mm. An exponential function could be used to describe the relationship between splash amount and raindrop diameter when raindrop accumulation number was 2000-4000 (P＜0.05). For all the raindrop diameters, the total splash amount presented an up-down-up trend as the particle size decreased. The amount of splashed fragments ＞2 mm was almost 0 for each raindrop diameter test. However, the amounts of splashed aggregates ＜0.053 mm increased with the increase of raindrop diameter. There was no significant difference in mean weight diameter of splashed fragments among different accumulation raindrop numbers or different raindrop diameters (P＞0.05). The variation in enrichment ratio with the fragment size was consistent under different raindrop diameters. Compared with the undisturbed soil, the splashed fragments in particle size >1 mm and < 0.053 mm were decreased for all raindrop diameters. However, the enrichment ratios of splashed fragments of the other particle sizes were greater than 1. The results can provide valuable information for agricultural development in the Loess Plateau.