Successful ecological restoration is expected to be accompanied by change in the propagule bank. We tested for temporal change in the soil propagule bank at two Bracken (Pteridium aquilinum)-infested sites in the United Kingdom (acid grassland and heathland), each with replicate experiments. A combination of bracken control (cutting, spraying, and combinations) and vegetation restoration treatments (seeding, fertilizer, harrowing) were applied. Soil propagule banks were sampled in 1998, 4–5 years after the start, and in 2003 after a further 5 years. We used univariate and multivariate statistical methods to investigate the response of the propagule bank to experimental treatment in space and time. Few effects were found in 1998, but after a further 5 years, the propagule bank size and composition changed significantly, implying that propagule bank development lags behind vegetation development. The effect of treatment on the propagule bank differed within sites. Thus, ecosystem development is occurring at different speeds and directions even in closely adjacent areas. Coupling between the propagule bank and vegetation changed between sampling times with either some new coupling or decoupling. At both sites, the propagule bank contained propagules of the target community and thus propagule bank development during restoration provides increased potential for vegetation recovery. However, it can take a considerable time for management effects to be detected in the propagule bank. Moreover, the effect or speed of effect is spatially variable. Continuing application of restoration treatments is recommended at heathland where there is a deep bracken litter layer.