The relevance of the oedometer tests used for the prediction of wetting-induced deformations in embankments is examined. Single and double oedometer tests are carried out. A comparison is made between laboratory tests and geotechnical centrifuge modelling at 100 g conducted to examine an inundated embankment made of a sand–clay mixture. A 20 cm high embankment model is built and instrumented. The material is compacted on the ‘dry side' of the optimum Proctor curve at a low compaction rate in order to emphasise settlement phenomena. The inundation simulation is conducted in two successive sequences during centrifuge flight up to a water table of 5 cm. The results prove that the prediction of the dry density after settlement due to inundation is good. The relevance of the oedometer tests used for the prediction of wetting-induced deformations in embankments is examined. Single and double oedometer tests are carried out. A comparison is made between laboratory tests and geotechnical centrifuge modelling at 100 g conducted to examine an inundated embankment made of a sand–clay mixture. A 20 cm high embankment model is built and instrumented. The material is compacted on the ‘dry side' of the optimum Proctor curve at a low compaction rate in order to emphasise settlement phenomena. The inundation simulation is conducted in two successive sequences during centrifuge flight up to a water table of 5 cm. The results prove that the prediction of the dry density after settlement due to inundation is good.