In recent years, nitrate contamination of groundwater has become an increasing concern for the health of people in rural areas who consume groundwater daily in China, ecially in the North China Plain.Usually, a deterministic model with mean hydraulic parameters is used to estimate the nitrate leaching for the whole agricultural field.In fact, the quantities of nitrate leached are variable from place to place in the field because of the parameters are different, so the simulated results by the deterministic model using mean hydraulic parameters could not represent the field-wide nitrate leaching.The objective of this paper is to evaluate the effect of spatial variability of the surface soil saturated hydraulic conductivity (K_s) on nitrate leaching and to assess the risk of nitrate contamination.First, a soil water-heat-nitrogen simulation model was established and verified based on the field data.Second, a sequential Gaussian simulation method was used to get a random field of the surface soil K_s based on 117 observed values in the field.Finally, the joint model was coupled with the random field of the surface soil K_s, which was gained by a conditional simulation method.During the period from March 15 to June 10, 1999, nitrate leaching was stochastically simulated with the irrigation rate being 353.8 mm and the nitrogen fertilizer application rate being 86.25 kg hm^-2.The accumulated quantities of the leached nitrate ranged from N 15.04 kg hm^-2 to 26.04 kg hm^-2 in the soil profile 2 m in depth, account ing for about 17.4% and 30.2% of the fert ilizer used during this period.The area with the probability of nitrate leaching exceeding the value obtained by the conventional method in the control being higher than 70% amounted to about 20%.It was implied that the soil spatial variability could result in change in nitrate leaching.The value obtained with the traditional method in the control could not be used because it obscured the risk of nitrate leaching.