Abstract: Increasing chemical fertilizer recovery is the best method of controlling N and P nonpoint pollution in vegetable fields. A 3-year orientation field experiment was carried out in kidney bean (Phaseolus vulgaris)-gourd ladle melon (Lagenaria sicerariae)-early rice rotating system in Pinghe County. The results show that the economic application rates of N and P for kidney bean are N 155 kg/ha and P2O5 79 kg/ha, and N 247 kg/ha and P2O5 130 kg/ha for gourd ladle melon,but no fertilization for early rice in the rotating system. Under the optimum benefit treatment, the net income is increased by 1.6%-46.8% compared with other application rates of N or P. In the treatment, the N fertilizer recovery is increased from 45.1% in the kidney bean-gourd ladle melon continuous cropping to 65.6% in the kidney bean-gourd ladle melon-paddy rice rotation, and the P fertilizer recovery is increased from 17.9% to 26.5%. The soil nitrate N and available P contents in the kidney bean-gourd ladle melon-early rice rotation are about 27.5% and 87.0% of those in the kidney bean-gourd ladle melon continuous cropping when the gourd ladle melon and paddy rice are harvested respectively, and are about 63.7% and 93.9% of those in base soils. Under the balanced fertilization, the nutritive peculiarity analysis shows that maximum uptake rates of N and P of kidney bean are 25.1 and 2.5 times of those of early rice, and are 18.4 and 1.2 times of those of gourd ladle melon respectively. The maximum uptake amounts of N and P of kidney bean are 1.7 and 2.2 times of those of early rice, and are 2.5 and 1.9 times of those of gourd ladle melon. The N and P uptakes are different in vegetable and rice, compared with vegetable continuous cropping system, the recoveries of N and P are enhanced in the vegetable-paddy rice rotating system, and the exorbitant enrichments and runoff amounts of N and P are reduced.