Abstract: The seed bank of seagrasses shows highly temporal and spatial heterogeneity and the variability can be as high as five orders of magnitude among seagrass species, ranging from 10⁰ to 10⁵. The soil seed bank of same seagrass species is also highly variable across habitats and seasons. Only those seagrass species with dormant seeds can form persistent seed bank (seed longevity≥1 a) such as Halophila,Syringodium,Ruppia and Cymodocea,or transient seed bank (seed-longevity <1 a) such as Zostera and Phyllospadix; while those seagrasses without distinct seed-dormancy (Enhalus and Thalassia), and viviparous seagrasses (Thalassodendron and Amphibolis) can not form soil seed bank. Vegetation characteristics, environmental conditions and marine animal behaviors, etc., can affect the seagrass seed bank. Two general methods, direct counting method and germination method (enumeration of germinated seedlings) are applied for estimating the size of the seed bank. The choice of method depends upon seed size and the ease with which dormancy can be broken. Maintaining the existing soil seed bank, especially the persistent seed bank, is crucial for the recovery of disturbed seagrasses, especially for those annual forms. Since most of seagrass meadows in China face intensive anthropogenic threats and many of them are composed of annual seagrasses, it is pivotal to conserve their seed bank for the recovery of these disturbed meadows.