Abstract We compared nekton and benthos densities and community compositions in a natural mixed seagrass bed dominated by Halodule wrightii (shoalgrass) with those found in three shoalgrass transplant sites and adjoining sand habitats in western Galveston Bay, Texas, U.S.A. Quantitative drop traps and cores were used to compare communities up to seven times over 36 months post-transplant where transplant beds survived. Total densities of fishes, decapods, annelids, benthic crustaceans, and most dominant species were significantly higher in natural seagrass than in transplanted shoalgrass or sand habitats during most sampling periods. On occasion, fish and decapod densities were significantly higher in transplanted shoalgrass than in adjoining sand habitats. No consistent faunal differences were found among transplant sites before two of three sites failed. Taxonomic comparison of community compositions indicated that nekton and benthos communities in natural seagrass beds were usually distinct from those in transplanted beds or sand habitats, which were similar. We conclude that reestablishing a shoalgrass bed that resembles a natural seagrass bed and its faunal communities in the Galveston Bay system will take longer than 3 years, provided that transplants persist.