Ship manoeuvring in shallow waters is more difficult and hazardous. In order to estimate the manoeuvring characteristics of the ship in finite water depth, the most effective and reliable method is to carry out the experiments in a shallow water towing tank. Generally, the ship model tests in limited water depth are achieved by varying the water depth in the towing tank. For large size towing tanks, primarily aiming at deep water manoeuvring , such as the 300m×18m×6m tank at the Hamburg Shipbuilding Research Institute (HSVA) , changing the water depth can be achieved either by the installation of an artificial bottom or draining the water out of the towing tank, which is time-consuming and inefficient. In order to improve the efficiency of ship model tests in shallow water, many towing tanks are equipped with the false bottom devices. Thus, varying the water depth can be achieved by adjusting the false bottom. However, a large full-size false bottom is usually accompanied with a small stiffness and will be easily deformed. Therefore, during shallow water tests, the false bottom is usually truncated at a limited size. This leads to the question on how much error could be introduced on the ship behaviour by differences in size and configuration of the false bottom. To the authors’ best knowledge, there is limited research that could quantify the error caused by a false bottom with limited horizontal dimensions. The objective of the present study is to investigate the interaction effects between the false bottom and the ship model, and to quantify the test errors (in percentage) due to different configurations and sizes of the false bottom. In order to achieve this, a 3-D Rankine source method based on the potential flow theory with linear free-surface condition is used to investigate the hydrodynamic interaction between the ship model and the false bottom. A contour of errors (in percentage) induced by the limited size of the false bottom was obtained, indicating the effects of the width and submerged depth of the false bottom.