Plant communities of tidal freshwater marshes fluctuate in composition seasonally and among years, but the influence of changes in hydrology on vegetation of these systems has not been examined. We investigated the effects of hydrology on vegetation of tidal freshwater marshes along the Patuxent River in Maryland, USA with a two-year field experiment, a one-year greenhouse experiment, and a seed-bank experiment. In the field experiment, sections of marsh soil and vegetation (“sods”) were elevated 10 cm, lowered 10 cm, or placed level with the marsh surface to simulate different hydrologic regimes. In the greenhouse experiment, sods were raised 10 cm above water surface (nonflooded) or flooded by 10 cm of water in tanks continuously, and in two other treatments changed from nonflooded to flooded or vice-versa after the first 35 days. For the seed-bank experiment, soil samples were spread in a 1-cm-thick layer in pans subjected to flooding by 3.5–4 cm of water or nonflooded but moist conditions in the greenhouse, and emerging seedlings counted. We found that lowering marsh sods by 10 cm (i.e., wetter conditions) in the field reduced plant species richness by 26% compared to the sods placed level with the marsh surface, while raising sods by 10 cm (drier conditions) increased richness by 42%. Total stem length of a majority of the most common species, as well as for all species combined, was more than twice as great in raised sods as in lowered sods. We observed similar patterns in richness and total stem length in the greenhouse study, where continuously nonflooded sods had almost twice the richness and 55% greater total stem length as continuously flooded sods. Sods that were flooded initially and then shifted to nonflooded conditions had richness and total stem length similar to the continuously flooded sods, while sods that were nonflooded initially and then flooded had richness and total stem length intermediate to continuously nonflooded and continuously flooded sods. In the field and greenhouse studies, species that are annual or annnual/perennial were more inhibited by flooding than were perennials. In the seed-bank experiment, flooding reduced the number of species emerging by 50% and total densities of emerging seedlings by 80% compared to nonflooded conditions. Taken together, the results of the field, greenhouse, and seed-bank studies indicate that 3–10 cm of flooding can significantly reduce seedling recruitment and growth in many plant species of tidal freshwater marshes and result in lower plant diversity. The greenhouse study further indicates that shallow flooding early in the growing season can reduce the abundance of certain species, primarily annuals, for the remainder of the growing season, resulting in a less diverse community. These findings suggest that hydrology is a dominant environmental variable controlling interannual variation in plant species composition of tidal freshwater marshes. Additionally, this study suggests that small increases in frequency and duration of inundation, which might occur due to watershed land-use changes, sea-level rise, or land subsidence, will reduce the diversity of these plant communities.