Salt marshes restored through managed coastal realignment (MR) often develop slowly and show persistent differences in vegetation from natural marshes. Development might be constrained by the availability of propagules or poor suitability of the abiotic environment for their establishment. To distinguish between these factors, we compared vegetation colonization and environmental conditions at a salt marsh created by MR at Brancaster, Norfolk, UK, with five reference marshes, varying in age from 30 to circa 6,000 years. After 5 years, plant communities of the MR site remained different from those in mature reference marshes. In contrast, the communities of the youngest reference marsh were not significantly different from mature reference marshes. At the MR site, abundance of perennial and later-successional species was low and large areas remained unvegetated. These differences are unlikely to be due to dispersal limitation, because 76% of the species from the local species pool colonized the site within 5 years. Although the annuals Salicornia europaea and Suaeda maritima were abundant by year 2, they were not ubiquitous until the end of the study. Tidal elevations of the MR site were suitable for vegetation development, but soil redox potentials were lower than that at the reference sites. Reducing conditions in the MR site appear to be the major cause of vegetation differences from the reference marshes, as they are associated with an abundance of bare ground and a small range of vegetation clusters. Measures to avoid low sediment redox potentials may have a great benefit in some salt marsh restoration projects.