Giant reed (Arundo donax) and Miscanthus spp. were tested to evaluate their tolerance and phytoremediation capacity in soils contaminated with heavy metals. Giant reed was tested under 450 and 900 mg Zn kg−1, 300 and 600 mg Cr kg−1, and 450 and 900 mg Pb kg−1 contaminated soils, while the Miscanthus genotypes M. × giganteus, M. sinensis, and M. floridulus were tested on 450 and 900 mg Zn kg−1 contaminated soils, along 2 years. Giant reed biomass production was negatively affected by the contamination; however, yield reduction was only significant under 600 mg Cr kg−1 soil. Zn contamination reduced significantly M. × giganteus production but not M. sinensis or M. floridulus yields. Yet, M. × giganteus was also the most productive. Both grasses can be considered as indicators, once metal concentration in the biomass reflected soil metal concentration. Regarding giant reed experiments, higher modified bioconcentration factors (mBCFs, 0.3–0.6) and translocation factors (TFs, 1.0–1.1) were obtained for Zn, in the contaminated soils, followed by Cr (mBCFs, 0.2–0.4, belowground organs; TFs, 0.2–0.4) and Pb (mBCFs, 0.06–0.07, belowground organs; TFs, 0.2–0.4). Metal accumulation also followed the same pattern Zn > Cr > Pb. Miscanthus genotypes showed different phytoremediation potential facing similar soil conditions. mBCFs (0.3–0.9) and TFs (0.7–1.5) were similar among species, but highest zinc accumulation was observed with M. × giganteus due to the higher biomass production. Giant reed and M. × giganteus can be considered as interesting candidates for Zn phytoextraction, favored by the metal accumulation observed and the high biomass produced. A. donax and Miscanthus genotypes showed to be well suited for phytostabilization of heavy metal contamination as these grasses prevented the leaching of heavy metal and groundwater contamination.