Arsenic (As) contaminated gold mine waste represents a significant issue for ecosystem and human health, as it is a known carcinogen and environmental toxin. Bendigo in Victoria, Australia, has reported As concentrations up to 22,000 mg/kg, exceeding the national soil guidance values (100 mg/kg) by 220-fold. To reduce exposure and potential human health risk, remediation or risk management strategies are required. Traditional remediation, such as excavation and landfill, often fail to improve soil conditions and are expensive. Biological remediation using plants (phytoremediation) represents a cost-effective strategy that results in the restoration of soil function and ecosystem services. There is a need to identify native plants for phytoremediation as current research focuses on fern species, which are not suitable for Bendigo. This study aimed to identify a native phytostabilising plant by conducting a mesocosm experiment involving Juncus usitatus, Poa labillardieri, and Themeda triandra. Plants grew in mine waste for 100 days; all survived and bioaccumulated As in the roots (phytostabilisation). Poa labillardieri grew significantly more biomass (roots 0.24 ± 0.01 g; shoots 0.22 ± 0.03 g (d/w)) and bioaccumulated the most As (~ 100 mg/kg plant biomass). The bacterial community was investigated to assess soil health and As association. The rhizosphere microbiota stabilised after root colonisation for Poa labillardieri and Themeda triandra, as indicated by richness and evenness changes. The two dominant phyla were Actinobacteriota and Proteobacteria, which displayed As tolerance. A keystone taxon, Latescibacterota, exhibited a positive relative association with As remediation. Poa labillardieri is a good candidate for future phytostabilisation trials.