Understanding these associations has important implications for ecological theory and for restoration practice. Here, we tested whether the presence of AMF and rhizobia influences the performance of native woody plants invaded by a non‐native grass in experimental microcosms. We planted eight plant species (i.e., Acacia acuminata, A. microbotrya, Eucalyptus loxophleba subsp. loxophleba, E. astringens, Calothamnus quadrifidus, Callistemon phoeniceus, Hakea lissocarpha and H. prostrata) in microcosms of field‐conditioned soil with and without addition of AMF and rhizobia in a fully factorial experimental design. After seedling establishment, we seeded half the microcosms with an invasive grass Bromus diandrus. We measured shoot and root biomass of native plants and Bromus, and on roots, the percentage colonization by AMF, number of rhizobia‐forming nodules and number of proteaceous root clusters. We found no effect of plant root symbionts or Bromus addition on performance of myrtaceous, and as predicted, proteaceous species as they rely little or not at all on AMF and rhizobia. Soil treatments with AMF and rhizobia had a strong positive effect (i.e., larger biomass) on native legumes (A. microbotrya and A. acuminata). However, the beneficial effect of root symbionts on legumes became negative (i.e., lower biomass and less nodules) if Bromus was present, especially for one legume, i.e., A. acuminata, suggesting a disruptive effect of the invader on the mutualism. We also found a stimulating effect of Bromus on root nodule production in A. microbotrya and AMF colonization in A. acuminata which could be indicative of legumes’ increased resource acquisition requirement, i.e., for nitrogen and phosphorus, respectively, in response to the Bromus addition. We have demonstrated the importance of measuring belowground effects because the aboveground effects gave limited indication of the effects occurring belowground.