The classification performance of a laboratory-scale teetered bed separator (TBS) was investigated through Whiten's efficiency curve approach. Higher levels of teeter water rate (TW) and pressure set point (SP) enhance the sharpness parameter (alpha) but reduce the fish-hook parameter (beta). The sharpness parameter (alpha) was found to be dependent on TW and SP. Apart from the process variables, density of the mineral has significant effect on the model parameters. It is established that the Whiten's efficiency curve approach has a potential to predicting classification performance of TBS by using experimental results with considerable degree of accuracy.