To quantify the impact of hydrogeological conditions on the migration distance of groundwater pollutants at unlined MSW landfill sites, this study established a 3D site-model for numerical simulation based on generalized hydrogeological model for plain-type landfill sites and groundwater pollution characteristics of unlined MSW landfills in China. Through numerical simulations with orthogonal experimental design, the 50-year migration distance of groundwater pollutants at three hydrogeological conditions in plain-type unlined MSW landfills were calculated. Range and variance analysis were conducted, and investigation results of typical site were compared with the numerical results. A BP neural network prediction model for 50-year migration distance was developed based on five hydrogeological parameters. The results of numerical simulations showed that the key factors influencing the 50-year migration distance of groundwater pollutants at unlined landfill sites were the permeability coefficient of the underlying layer, the relative depth of the water table at the landfill site, as well as the leachate table within the waste pile. The BP neural network prediction model reached a coefficient of determination of 0.87, and the prediction of the landfill site case was consistent with the simulated and measured results, indicating that the 50-year migration distance of groundwater pollutants at unlined landfill could be predicted using the five key hydrogeological parameters. The study provided guidance for determining the boundaries of groundwater environmental investigations at plain-type MSW landfill sites.