Direct current electrical resistivity method has experienced significant breakthroughs during the last decades with the development of advanced instrumentation and sophisticated inversion algorithms. These have substantially benefitted the field of archaeological prospection to extract quantitative information for the buried archaeological material in a complete three-dimensional (3-D) context. The obvious continuation of past human occupation and associated settlements in the ultra-shallow part of coastal zones generated the necessity to compile methodologies for mapping submerged cultural assets. This study investigates the efficiency of dynamic floating and submerged 3-D Electrical Resistivity Tomography (ERT) for mapping archaeological relics buried beneath the sediments in ultra-shallow marine environments. Extensive testing is performed through numerical simulation and 3-D inversion of synthetic apparent resistivity tomographic data. The generic resistivity model including a complex resistive structure embedded in a stratified conductive environmental regime simulated typical scenario of an archaeological feature submerged below the sea water layer. Different array configurations including Dipole-Dipole (DD), Gradient (GRD), Reciprocal Wenner (RecWEN) and Wenner-Schlumberger (WS), which can be appropriately programmed for continuous off-shore measurements with associated multichannel resistivity instruments, are validated in order to determine the most efficient one for such surveys. Densely spaced multiple parallel two-dimensional (2-D) ERT lines along a single direction composed the survey layout to extract the 3-D apparent resistivity data set for the different electrode arrays using a 3-D finite element program. The synthetic tomographic data were corrupted with 3% Gaussian noise and inverted with an iterative smoothness constrained inversion algorithm. The comparative results from the various tested arrays manifest the superiority of the DD in reconstructing the optimum resistivity inversion model for both the floating and submerged survey modes. Additional tests were made concerning the resolving capabilities of ERT with variable seawater thickness and target characteristics.Although accurate 'a priori' information regarding the water resistivity and thickness are essential for constraining the 3-D inversion, erroneous estimation of these parameters can result to misleading results, especially for submerged survey modes.The simulation of floating and submerged 3-D ERT surveys through synthetic modelling documented the benefits of such approach in reconstructing structured cultural