Abstract: In order to reveal the different effects and benefits of reverse-slope terrace on controlling nitrogen and phosphorus loss on sloping farmland, the loss of runoff, sediment, nitrogen and phosphorus after rainfall was observed in 2 sloping farmland standard runoff plots (1# and 2#) in central Yunnan Province from 2011 to 2016. The 1# plot is the original sloping farmland, and the 2# plot is the sloping farmland with reverse-slope terrace. The effects of early rainfall were eliminated; finally the data of 45 typical rainfalls were selected. Based on these data, 4 rainfall patterns were classified according to the amount, intensity and erosivity of rainfall with 2 methods including quick clustering and discriminant clustering: Type A (great amount, strong intensity and high erosivity of rainfall), Type B (low amount, minor intensity, small erosivity of rainfall), Type C (low amount, low medium intensity and erosivity of rainfall), Type D (high medium amount, intensity and erosivity of rainfall). Type B was the most frequent rainfall type in the study area, but soil and water loss in the red soil sloping farmland was mainly caused by Type A and Type D. Moreover, the reduction effect for runoff and sediment of reverse-slope terrace was extremely significant (P<0.01), the reduction rate of runoff was 52.11% and the reduction rate of sediment was 71.30%, and the reduction rate of sediment was significantly greater than that of runoff (P<0.01); the reduction rate of runoff was presented as Type C > Type B > Type D > Type A, and the reduction rate of sediment was presented as Type D > Type A > Type C > Type B; the reduction rate of runoff by reverse-slope terrace in Type C and Type B was significantly higher than that of Type D and Type A (P<0.05), while the sediment reduction rate was significantly lower than that of Type D and Type A (P<0.05). The reduction rates of reverse-slope terrace for total nitrogen, NO3?-N, NH4+-N in runoff were 68.10%, 69.81% and 69.81%, respectively, and the reduction rates of total phosphorus, dissolved inorganic phosphate (DIP )in runoff were 71.52% and 72.77%, respectively. It was presented that the reduction rates of reverse-slope terrace for total nitrogen, NO3?-N, NH4+-N, total phosphorus, and DIP in runoff were reduced with the increase of the rainfall amount and intensity in different rainfall patterns. In addition, reverse-slope terrace could cut 57.32% of total nitrogen, 54.22% of hydrolyzable nitrogen (HN), 67.38% of total phosphorus total phosphorus and 63.69% of available phosphorus (AP), and the reduction rates of total nitrogen, HN, total phosphorus, and AP in sediment were increased as the rainfall amount and intensity rose under different rainfall patterns. In conclusion, this study is very helpful both in revealing the mechanism and evaluating the efficiency of reverse-slope terrace to control non-point source pollution in sloping farmland. It also provides technical support for water and soil conservation and controlling agricultural non-point source pollution in mountainous areas from the source.