Nanoqueous phase liquid (NAPL) simulator is a powerful and popular mathematical model for modeling the flow and transport of non-aqueous phase liquids in subsurface, but the testing of its feasibility under water table fluctuation has received insufficient attention. The feature in a column test was tested through two cycles of water table fluctuation. The sandy medium in the column was initially saturated, and each cycle of water table fluctuation consisted of one water table falling and one rising, resulting in a drainage and an imbibition of the medium, respectively. It was found that the difference between the simulated and measured results in the first drainage of the column test was minor. However, with the propagation of the water table fluctuations, the simulation errors increased, and the simulation accuracy was not acceptable except for the first drainage in the two fluctuation cycles. The main reason was proved to be the estimation method of residual saturation used in this simulator. Also, based on the column tests, it was assumed that the resulting residual saturation from an incomplete imbibition process was a constant, with a value equal to that of the residual value resulting from the main imbibition process. The results obtained after modifying NAPL simulator with this assumption were found to be more accurate in the first cycle of water table fluctuation, but this accuracy decreased rapidly in the second one. It is concluded that NAPL simulator is not adequate in the case of LNAPL migration under water table fluctuation in sandy medium, unless a feasible assumption to estimate residual saturation is put forward.