The observations from 14-yr long-term investigation on the soil-water losses in the sloping red-earth (slope 8°- 15°) showed that soil-water losses were closely correlated with land slope and vegetative coverage. Runoff rate in sloping red-earth could be reduced doubly by exploitation, while the soil erosion was enhanced doubly during the first two years after exploitation. Subsequently, it tended to be stable. Soil erosion was highly positively correlated with land slope, i. e. soil erosion increased by 120 t km-2 yr-1 with a slope increase of 1°. On the contrary, soil erosion was highly negatively correlated with vegetative coverage, i. e. soil erosion was limited at 200 t km-2 yr-1 below as the vegetative coverage exceeded 60%. Furthermore, soil erosion was highly related with planting patterns, i. e. soil erosion in contour cropping pattern would be one sixth of that in straight cropping. Based on the view of soil nutrient balance and test data, it was first suggested that the soil loss tolerance The observations from 14-yr long-term investigation on the soil-water losses in the sloping red-earth (slope 8 ° - 15 °) showed that soil-water losses were closely correlated with land slope and vegetative coverage. Runoff rate in sloping red-earth could be reduced doubly by exploitation, while the soil erosion was enhanced doubly during the first two years after exploitation. km-2 yr-1 with a slope increase of 1 °. On the contrary, soil erosion was highly negatively correlated with vegetative coverage, ie soil erosion was limited at 200 t km-2 yr-1 below as the vegetative coverage exceeded 60% Soil erosion was highly related with planting patterns, ie soil erosion in contour cropping patterns would be one sixth of that in straight cropping. Based on the view of soil nutrient balance and test data, it was first suggested that the s oil loss tolerance