<p>The embankment dams are mostly built using the fill (rock, and soil) materials and use their mass to overcome the forces such as sliding or overturning. Failures of such dams may involve structural, hydraulic, seepage and piping problems and one of the challenges of this in geotechnical engineering relates to the variable foundation conditions. To predict seepage in the Khasa Chi Dam, the artificial intelligence model of the adaptive neuro-fuzzy interference system, or ANFIS, and artificial neural network (ANN), was used for this investigation. In order to get piezometric head readings at three distinct points along the dam body, the upstream and downstream water levels were interpreted as the input with the exception of two piezometers, the findings showed good model performance. ANN and ANFIS models are powerful tools to assess dam safety, with a small number of field data (R² &gt; 0.98, NSE &gt; 0.98). They provide a reliable platform to predict seepage situation in earth-fill dams when such conditions belonging to various parameters are different, and ultimately it permits them to compare the performance. It used established statistical quantities, coefficient of determination (R²), root mean square error <b>(</b>RMSE) and Nash-Sutcliffe Efficiency (NSE) to give the findings. This research was useful in concluding that Single and multiple ANN and ANFIS models are likely to exhibit same predictive capabilities with ANN predicting better on certain piezometers albeit that ANN was simpler to build than ANFIS.</p>