Summary A three dimensional finite element model, based on Galerkin weighted residual techniques, is presented for groundwater simulation in the lower Indus Basin, Pakistan. The model was calibrated against field data collected at different agricultural farms located in the Khairpur district. Twenty six observation wells were installed to monitor the groundwater levels for model calibration. The values of the statistical performance parameters adjusted R2, mean absolute error (MAE), root mean square error (RMSE), Nash–Sutcliffe efficiency or model efficiency (ME), BIAS, and index of agreement (d) showed that the overall model performance for steady and transient groundwater flow is good. The calibrated model was used to assess the impacts of different well pumping rates, well screen lengths, and canals head boundaries on the extent of waterlogging. The model results suggest that well pumping rate is a prominent factor to control waterlogging. An increase in well pumping rate by 25% decreased the water logged area by 16%, while an increase in pumping rate by 50% decreased the water logged area by 25%. The waterlogging in the study area was attributed to the variations in canal water levels. It was further observed that waterlogged area with a watertable depth less than 0.8 m is increased by 5.8% when the water level in the Khaipur Feeder East (KFE) canal was increased by 0.6 m while the water level at the Rohri canal was kept constant, at the pumping rate of 1728 m3 d−1. Similarly, when the water level at the Rohri canal was increased by 1 m whilst that at the KFE was kept constant, the area under waterlogging had increased by 10.5%. If water levels in both canals were to be increased simultaneously (0.6 m in KFE and 1.0 m in Rohri canal) the waterlogged area will increase by 18.1% for the given well discharge.