A Point Count System Model, DRASTIC, of U.S. Environmental Protection Agency was used to assess the vulnerability of groundwater from non-point source pollution for Ludhiana district of Punjab, India. Thematic maps of seven parameters of the model i.e. depth of water table, recharge, aquifer media, soil, topography, impact of vadose zone and hydraulic conductivity were prepared in a geographic information system (GIS) environment to obtain pollution potential indexes based on the hydo-geological paramaters and consumption of nitrogenous fertilizers. The model was calibrated by using the data from first four years of the study period. In the calibration process, the assigned weights were modified to improve the correlation coefficient between model computed values and the observed values. The values of these improved weights for each parameter were used to calculate the regression equation for each block of the study area. These regression equations were then used to validate the model for the next four-years. Statistical analysis (Root Mean Square Error and Average Absolute Error) was carried out to compare the model output with the observed values at different grid cells and found good agreement between different statistical parameters. The modified weights can be used for future estimation of nitrate as well as any other pollutants of the groundwater in the areas of similar conditions.