A sampling protocol for the assessment of seed banks in Australian temporary and permanent wetlands is presented. A simple corer design allows removal of intact samples from water depths up to 1.2 m, ensuring comparable sampling in a variety of wetland habitats. Sampling and experimental designs appropriate for answering questions about the germination of seeds from sediments under different water regimes in both the glasshouse and the field are presented. The adequacy of area sampled and of number of replications for assessment of species richness and abundance of seeds in the seed bank is considered, and a minimum number of five aggregate samples (each sample of eight cores has an area of 0.016 m2) is suggested for these wetlands. The variation between samples in this study is within the range reported for seed-bank studies in similar wetlands in North America.