The actual pollutant transport concentration and pollution range are affected by the numerical uncertainty and random spatial distribution of hydrogeological parameters. Conventional numerical models cannot effectively simulate the transport of pollutants. So, we integrated the numerical uncertainty and random spatial distribution of hydrogeological parameters into the conventional numerical model respectively for better simulation results. Taking the chloride leachate transport of a landfill as an example, the variation of its concentration and range from 1940-2000 year is calculated. Taking the results of year 43 as example, the results show that the mean values of multiple simulations of the random models which takes into account the uncertainty factors were consistent with the actual values, indicating that the proposed models are superior to conventional numerical models. It is concluded that when the variation range of the independent variables was 20%, the mean value and mean variance values of numerical uncertainty random model were 25.04% and 26.95% larger than those of the random spatial distribution model. The uncertainty of calculation results of chloride concentration and pollution range varied in the study period, and the variation trend of the uncertainty of calculation results of two kinds of uncertainty models were consistent which increased first then decreased. Through uncertainty analysis, the variation range of chloride concentration in this area are clarified, the risk analysis can be carried out on the basis of dividing different concentration confidence intervals, so as to provide data reference for the prevention and control of groundwater pollution.