Abstract: Abstract: A small watershed is one of the most basic units in soil and water loss control. It is of great significance for the precise layout of the soil and water conservation in small watersheds. Soil erosion control is often required for the accurate identification of erosion and sediment yield areas, as well as sediment transport paths. Furthermore, the soil erosion and sediment connectivity can be coupled for the decision-making on the watershed sediment. However, this approach cannot be effectively applied at the scale of the small watershed. Therefore, this study aims to evaluate the erosion and sediment yield in small watersheds using GeoWEPP and sediment connectivity. The study area was then taken as the typical agricultural watershed in the upper reaches of the Lijiang River. The high-resolution images were collected from the UAV remote sensing. Firstly, the GeoWEPP model was used to determine the slope sub-area and soil erosion hotspots with the soil erosion risk in the small watershed. The simulation and evaluation of the erosion and sediment yield were carried out on the soil water erosion. The spatial pattern of erosion and sediment yield was clarified in the small watershed. Secondly, the monitoring device was installed at the outlet station of the small watershed. The runoff and sediment data were collected to adjust the parameters of the model. The sediment connectivity was then calculated to consider the combined effects of topography and vegetation, as well as topography, vegetation, roads, and ditches. Finally, the spatial distribution of sediment connectivity was obtained after calculation. As such, the optimal algorithm was selected to quantitatively analyze the sediment connectivity hotspots in small watersheds. The results show that: 1) The reliable simulation was found in the Jianlishui small watershed. The determination coefficients of runoff and sediment yield were 0.811 and 0.857, respectively, and the Nash-Sutcliffe deterministic coefficients NSE were 0.887 and 0.829, respectively. The average modulus of annual soil erosion was 1 092 t/ (km²·a), indicating mild erosion. The extremely high-risk area only accounted for 2.91% of the total area of the basin. However, a contribution of 13.63% was observed in the total sediment yield. 2) The average value of sediment connectivity index (IC) in Scenario 2 significantly increased after integrating into roads and ditches, compared with Scenario 1. The better performance was achieved to more accurately identify extremely low and extremely high connectivity areas, thus representing the real situation. The distribution of IC was similar to the trend of natural terrain, ponds, roads, and ditches, with higher connectivity near ponds, canals, and roads. 3) The contribution rate of erosion hotspots to sediment transport in small watersheds was greater than that of the connectivity hotspots. The overlapping areas of the high erosion and the high/extremely high connectivity accounted for only 0.83% of the total. However, a significant contribution to sediment transport was observed for soil erosion control. This finding can also provide a strong reference for the precise layout of the soil and water conservation in small watersheds.