Abstract: Abstract: It is of great importance to control the spatial distribution pattern and evolution tendency of soil erosion in the Dianchi Basin for optimizing soil and water conservation measures, conducting water pollution control of Dianchi Lake and ensuring the basin's sustainable development. In this study, the amount of soil erosion in the Dianchi Lake basin between 1999 and 2014 has been calculated every 3 years by applying RS and GIS techniques in combination with a soil erosion model (USLE) based on the data source of precipitation, soil, DEM and remote sensing images, for the purpose of analyzing the spatial-temporal evolution characteristics of the basin's soil erosion intensity. The results indicated that: 1) The USLE model showed a high simulation accuracy for the erosion amount with the relative error was ≥18.16% and R2≥0.996; 2) The area with no apparent erosion increased from 70% in1999 to 82% in 2014, while there was a year-by-year downtrend for the soil erosion area, with the erosion area in 1999 was 776 km2 and decreased to 468 km2 in 2014, which showed that soil erosion in the Dianchi Basin has been effectively controlled; 3) There were remarkable characteristics for the spatial evolution trend of basin's erosion intensity: the erosion intensity in 75% or so of the basin did not change, and was spatially concentrated in the main urban area of Kunming City, the downtown of Jinning County and the area of Xinjie Town; 18.23% of the basin showed a decline in the level of erosion intensity, mainly concentrated in the streams around Dianchi lake and flowing into the lake; 8.36% of the basin showed a rise in the level of erosion intensity, mostly distributed in the mountainous terrains at the east, north and south ends of the Dianchi Basin; 4) The erosion intensity level diverting matrix showed that a relatively small change rate and a relatively high preservation rate in tiny erosion; the rate of high erosion levels shifting to low erosion levels was higher than that of low erosion levels shifting to high erosion levels, illustrated that the erosion areas were well managed with significant effects; the rate of tiny erosion and slight erosion shifting to moderate and intensive erosion was very low, lower than 0.94% in all periods, which meant that the policy of converting farmland into forests launched since 1998 obtained quite good effects. In sum, the production of series of graphs for erosion changes by integrating RS, GIS techniques and the soil erosion model can help to discover vulnerable areas for soil erosion from a spatial-temporal change perspective, which could be helpful to analyze the evolution of regional soil erosion from emergence, development to disappearance, providing the bases for making decisions for further accurate erosion prevention/control measures and as a reference for the basin's eco-environmental protection.