Probabilistic analyses are conducted for seepage through a rock-fill dam having two different core-filter configurations: one sloping and the other a central symmetrical core-filter arrangement. Uncertainties in core and filter are considered, assuming their hydraulic conductivities as random variables. For this purpose, finite-element software used for groundwater flow and seepage analyses is coupled with a random-number-generation algorithm. Monte Carlo simulations are performed for probabilistic seepage analyses. The suggested method is applied to a 30-m-high rock-fill dam. Results showed that the uncertainty in core and filter causes a decrease in the seepage rates of both configurations. It is concluded that in case of uncertainty, the central symmetrical core-filter configuration is more successful than the sloping configuration in decreasing seepage.