It is agreed in most researches that vegetation coverage can effectively reduce soil erosion and this reduction is attributed to interception of rainfall by stems and leaves, fixation of soil by roots and resistance of vegetation covers to runoff movement. But because total soil nitrogen mainly exists in organic forms and combines with soil particles, vegetation covers, preventing soil particles from being eroded, reduce soil total nitrogen loss. The effect of vegetation covers on soil erosion will be intensified as the vegetation coverage is increased. Unlike total soil nitrogen, mineral soil nitrogen (NH + 4 N and NO - 3 N) mainly exists inside soil liquid and attaches itself to the surface of soil particles. The catchment model of 1∶400 scale under artificial rainfall conditions with rainfall intensity of 2 mm/min and rainfall amount of 60mm in 1998 and the original catchment with an area of 8 05 km² under natural rainfall conditions in 1992 and 1998 were utilized to study the relationship of the nitrogen loss of catchment flood with vegetation coverage and comprehensive control to erosion. This research is aimed at finding out the relationship between comprehensive catchment control and nitrogen loss by erosion that will serve as the scientific base for reseeding cropland to forest and grass. As the vegetation coverage was increased from 0 to 60%, the flow of runoff was reduced only by 19 9%; the time span from the end of rainfall to the occurrence of soil erosion was lengthened from 1 5 minutes to 10 2 minutes and multiplied 5 8 times as much; the concentration of mineral nitrogen in runoff was increased from 1 4 mg/kg to 4 5 mg/kg and multiplied 2 2 times as much. The results showed that as the vegetation coverage was increased in the catchment, the loss of mineral soil nitrogen was intensified. But further researches should be conducted as to the intensity and mechanism of the interaction between runoff and mineral topsoil nitrogen. On June 23 and August 21, 1998, the concentrations of NH + 4 N in rainwater were 1 4 mg/kg and 1 9 mg/kg and the concentrations of NO - 3 N in rainwater were 1 2 mg/kg and 1 7 mg/ kg respectively. On the corresponding period, those of NH + 4 N in river water were 1 2 mg/kg and 0 9 mg/kg respectively and those of NO - 3 N in river water 1 3 mg/kg and 0 6 mg/ kg respectively. Thus it can be seen that the catchment could function as a filter of the available rainfall nitrogen that retained more NO - 3 N than NH + 4 N. In the catchment, the loss of NH + 4 N was 6 525 kg/km² and 2 725 kg/km² respectively, the loss of NO - 3 N was 5 101 kg/km² and 1 258 kg/km² respectively, the loss of total nitrogen was 346 8 kg/km² and 42 7 kg/km² respectively, and the loss of organic matter was 5 201 4 kg/km² and 552 3 kg/km² respectively. The enrichment of micro aggregates with a diameter of less than 20 mm led to the enrichments of organic matter and total nitrogen in river sediment. Because of water and soil loss, the annual loss of total soil nitrogen content was 26 1%, 52 6%, 21 5%, 8 8% and 7 7% from 1992 to 1998 respectively in the lands for wood, economic forest, shrub, grass and crops. In the catchment, the erosion induced soil loss was 1 086 t/km² and 1 119 t/km² respectively. The erosion induced loss soil nitrogen was 8 758 5 kg and 7 562 2 kg respectively, which meant a decrease of 15 8% from1992 to 1998. The erosion induced loss soil nitrogen was decreased from 1 207 2 kg to 579 9 kg in the farm land in the catchment and thus the decreasing rate was 52 0% from 1992 to 1998. The erosion induced loss soil nitrogen in the farming land accounted for 13 8% and 7 7% of the total soil nitrogen loss in the catchment in 1992 and 1998.