Aquatic ecosystems are seriously threatened because of fragile Eco-environment and severe non-point pollution in Three Gorges Reservoir Area of China. Crop/mulberry intercropping is popular because of efficiency to reduce nitrogen (N) and phosphorus (P) loss and control agricultural non-point source pollution. However, there is a huge controversy to achieve optimal crop/mulberry intercropping in the Three Gorges Reservoir Area. In order to optimize the crop/mulberry intercropping pattern in purple dry slope-land so as to improve its effects of soil and water conservation and eco-environment benefits in the Three Gorges Reservoir Area, an experiment with five treatments (i.e. Treatment I: Contour tillage with contour mulberry tree lines on upper-slope, middle-slope and lower-slope, Treatment II: Cross cultivation with contour mulberry tree lines on upper-slope, middle-slope and lower-slope, Treatment III: Contour tillage with contour mulberry tree lines on middle-slope and lower-slopes, Treatment IV: Cross cultivation with contour mulberry tree lines on middle-slope and lower-slopes and Treatment V: Conventional contour tillage.) was carried out in Wangjiagou catchment of the Three Gorges Reservoir Area from January 1, 2014 to December 31, 2014. N and P in surface runoffs was analyzed for forms (ammonium nitrogen, nitrate nitrogen, total nitrogen (TN), dissolved nitrogen (DN), particulate nitrogen (PN) and total phosphorus (TP), dissolved phosphorus (DP), particulate phosphorus (PP)) and annual loading of N and P loss with surface runoff was assessed. Results showed that on the whole, crop/mulberry intercropping significantly retarded N and P losses with surface runoffs, but the effect varied sharply with rainfall intensity and pattern of the intercropping. Treatment II was the most effective in retarding the loss of TN, DN, PN, TP, DP and PP, regardless of rainfall intensity and Treatment III came the next. However, under rainfalls moderate or low in intensity (44.2 mm d^-1 and 53.9 mm d^-1), Treatment I was the most effective in retarding the loss of ammonia nitrogen and nitrate nitrogen, while under rainfalls high in intensity (122.4 mm d^-1), Treatment I was obviously lower than Treatment II and Treatment III in such an effect. Furthermore, the treatments also differed sharply in annual loading of N and P loss with surface runoff because the effects of cultivation practice, location of contour mulberry tree lines, and their interactions on N and P losses were different. Compared with Treatment V, Treatment III reduced TN by 12.8%, TP by 19.3%, DN by 20.2%, DP by 10.9% and PP by 25.7%; Treatment IV reduced TN by 41.7%, TP by 44.7%, DN by 44.7%, DP by 38.1%, PN by 36.1% and PP by 48.5%; Treatment I reduced TN by 23.7%, TP by 42.5%, DN by 25.5%, DP by 40.2% , PN by 19.7% and PP by 44.1%; and Treatment II reduced TN by 43.6%, TP by 58.8%, DN by 45.7%, DP by 55.4% , PN by 39.3% and PP by 61.1%. The nitrogen and phosphorus lost with runoff were dominated with DN (approximately 50.1% ~ 60.2%) and PP (approximately 54.9% ~ 59.6%), and the annual loading of NO₃⁻-N loss (approximately 0.19 ~ 0.27 kg hm^-2 a^-1) was higher than that of NH₄⁺-N loss (approximately 0.12 ~ 0.17 kg hm^-2 a^-1). In conclusion, because of combined effects of cultivation practice, location of contour mulberry tree lines, and their interactions on surface N and P losses, four kinds of crop/mulberry intercropping practices reduced surface runoffs, N and P concentration in surface runoff caused by each rainfall event, and N and P annual loading. But treatment II (cross cultivation with contour mulberry tree lines on upper-slope, middle-slope and lower-slope) is more pronounced in retarding and controlling N and P loss with surface runoff, and therefore it is an optimal intercropping pattern for exploiting the resources of purple dry slope-land in the Three Gorges Reservoir Area of China.