Transitions between aquatic and terrestrial ecosystems represent zones where soil moisture is a dominant factor influencing vegetation composition. Niche models based on hydrological and vegetation observations can be powerful tools for guiding management of these zones, especially when they are linked with physically based hydrological models. Floodplain restoration represents a unique opportunity to utilize such a predictive vegetation tool when a site's hydrology is altered to create a wetter environment. A variably saturated groundwater flow model was developed and used to simulate the soil moisture regime across a floodplain in Wisconsin where post-settlement alluvium was removed with the intent of increasing regionally threatened wetland plant species. Hydrological niche models based on simultaneous observations of vegetation composition and surface effective saturation were used to predict probability of presence for two plant species (Carex vulpinoidea (fox sedge) and Elymus canadensis(Canada wildrye)) and wetland indicator score (a composite indicator of relative frequency of species in five habitat categories) based on simulated surface effective saturation. The vegetation predictions following restoration are more wetland-species dominant overall. However, zones of the study site where a confining layer is present that decouples groundwater from the near-surface soil zone tend to be drier following restoration due to restricted upward groundwater flow and less soil water storage above the confining layer. As reflected by an increase in the interquartile range in the predicted wetland indicator score, this restoration technique may increase the site-scale spatial diversity of plant community types while simultaneously accomplishing the goal of increasing wetland plant species occurrence.