After decades of exploitation of hydrocarbons, the offshore facilities constructed for this purpose are nearing the end of their design life and/or economic operation. According to international law, these offshore structures need to be completely removed. Decommissioning of the substructures often requires for soil to be removed in piles to facilitate pile cutting works below the seabed. One way of achieving this is by deploying a specialized tool, a so-called Soil Plug Removal Tool (SPRT), that operates using hydraulic excavation. This research carried out under supervision of Royal Boskalis Westminster N.V. aims to get a more solid basis for comparison of different SPRT concepts that are available in the market. The tools are designed to handle a wide range of soil types. Removal of cohesive sediment is more challenging, mainly due to the very low water permeability present compared to granular soils. This study therefore focusses on the excavation of cohesive soil types only. In order to verify the performance of several concepts an experimental test program is set up on model scale. The primary goal is to investigate the achievable excavation production in terms of tool progress rate. Therefore, a jetting tool is developed that covers the (complete) spectrum in terms of cohesive soils and performance of the available tools. Two basic SPRT concepts are incorporated in this single tool based on head movement: static or rotating. Jet pressure, clay strength, rotational velocity and set down pressure of the tool are altered during testing on the condition that all other parameters are fixed. This requirement is met for the testing clay by merely varying the shear strength. The testing clay was therefore prepared both with an artificial and natural clay with shear strengths ranging from 20 kPa to 100 kPa. It is found that next to jetting, soil failure can also be attributed to cutting and jet trench failure under influence of the jetting head that rests on top of the clay. For this reason, production values belonging to jetting could not be obtained directly and had to be calculated using jetting theory to distinguish between jetting and jet trench failure. Based on the power that is required to excavate a certain volume of soil (i.e. specific energy), insight is given in the contribution of each failure mechanisms to the production in terms of tool progress rate. During static jetting, the current (nozzle) configuration did not remove enough soil from the jet cavities for the jetting tool to progress downwards. The opposite is true for the rotational tests which comprise the largest part of the test series. An analytical model is proposed to predict the cavity width and depth. This model is only valid for jets with small rotational velocities as encountered in this study. The total production, which was measured, is found to be inversely proportional to shear strength and directly proportional to jet pressure and rotational velocity.