Abstract: 【Objectives】The study of the effects of long-term returning of crop residues and fertilization oncrop yields and soil fertility can offer the scientific base for the best nutrient management in cropping system. 【Methods】 A long-term field experiment with rice-rapessed rotation has been conducted in Guanghan City, Chengdu Plain since 2005. The experiment consisted of eight treatments: chemical fertilizers without straw mulch (CK), four rates of nitrogen, two rates of phosphorus and three rates of potassium under straw mulch, respectively. Since 2005, plant samples of each plot were collected and crop seed/grain yields were recorded at harvest, soil samples were collected at 0-5, 5-15 and 15-30 cm depths of each plot after rice harvested in 2010. 【Results】 Under rice-rapeseed rotation, crop yields vary considerably from year to year due to the weather changes. Rice yields were significantly affected by the surface mulched rapeseed residue since the third year of the experiment. The rice yield of the treatments with the straw mulch was significantly higher than those treatments receiving equal amount of N fertilizer without mulch, and equivalent or even higher rice yields were obtained in treatments of -P and -K with mulch than the CK. Thus, about 60 kg of N and 90 kg of K2O per hectare could be cut off from the fertilizer input through rapeseed residue mulch in the rice season. In the rapeseed season, however, rice straw mulch was not able to reduce nitrogen input, the -P treatment produced the lowest rapeseed yield, indicating that P fertilizers should be applied in the rapeseed season, but less or not in the rice season. In the rapeseed season, about 30 kg of N and 90 kg of K2O per hectare can be saved by rice straw mulch. Continued straw mulch and no-tillage mainly benefit the accumulation of soil organic matter, N, P and K at 0-5 cm soil layer, much less in the soil layers below. The -P and -K treatments obviously depleted the soil P and K reserves and the CK treatment mainly depleted soil K reserves. Under the same P and K rates, contents of soil available K and P were decreased with the increase of N input rates. Long-term straw mulch and no-tillage dramatically decreased soil bulk density, increased soil porosity, non water-stable aggregates and water-stable aggregates, improved the physical properties of the 0-5 cm soil layer.【Conclusions】Long-term straw mulch plus no-tillage in the rice-rapeseed rotation system could significantly increase crop yields, reduce fertilizer inputs (particularly in rice season) and improve top layer soil fertility. Adequate P fertilizer should be applied in rapeseed, less or not in rice for high yield and high profit production.