Biodiversity loss is a global crisis, due primarily to habitat destruction and widespread nonnative invasions. Invasive grasses are particularly problematic in many tropical ecosystems, where they possess traits that promote their persistence and can drastically alter native plant communities. We explored the ecophysiological basis for restoring native Hawaiian dryland ecosystems currently dominated by the nonnative invasive grass Megathyrsus maximus (guinea grass) in a garden experiment. Three native species—Myoporum sandwicense (naio; canopy tree), Dodonaea viscosa (aalii; shrub), and Plumbago zeylanica (iliee; groundcover)—were grown with M. maximus at three levels of native functional diversity (one, two, or three species) while holding overall plant density constant. We tested which individual and functional combinations of native species were more productive and best suppressed M. maximus growth and reproduction. Megathyrsus maximus had 39–94% higher maximum photosynthetic rates (A_max) than native species and increasing native functional diversity did not affect M. maximus A_max. Aboveground, belowground, and total biomass of M. maximus varied with functional diversity, although intraspecific competition reduced growth as much as interspecific competition. Reproductive tiller production by M. maximus decreased significantly when planted with any of the native species and with increasing native functional diversity. These results indicate that high native functional diversity in an ecological restoration setting may aid in the control of a dominant invasive grass and the reintroduction of diverse native species. Recommendations for restoring degraded nonnative grasslands in Hawaii and throughout the tropics include selection of native species that are ecophysiologically competitive and have high functional diversity.