There is little experimental evidence to guide the design of substrate compaction for the optimal plant water use performance of a landfill phytocap. A glasshouse study was undertaken to address this with aim to evaluate the effect of substrate compaction on the water use of a phytocap plant community. Four levels of relative compaction (RC), i.e. the ratio of dry bulk density over the standard maximum dry bulk density, (72%, 77%, 82%, and 87%) were considered. The native tree and grass species selected were typical of an Australian phytocap plant community: Themeda triandra, Microlaena stipoides, Eucalyptus camaldulensis, Eucalyptus cladocalyx, Acacia mearnsii and Allocasuarina verticillata. Plant water use was measured by weight as the difference between planted and unplanted cores over 5 drying periods occurring through the first 6 months of plant establishment traversing winter, spring and summer. Plant water use was optimal for all species at low-intermediate RC (72%, 77% and 82%), and all species except Themeda triandra, were most negatively impacted by the highest RC of 87%. The best linear model based on Akaike’s Information Criterion included a second-order term for the continuous fixed factor ‘RC’ and the categorical fixed factor ‘species’. This model showed plant water use to be optimum at a RC of 76.5% and highlighted a wide range of RC’s (70–83%), for which plant water use is not <90% of this optimum. It also highlighted increasing plant water use-sensitivity to RC’s beyond these ranges, with a RC > 86% and a RC < 67% leading to reductions in plant water use of 20% or more. Substrate specifications are recommended to optimize phytocap plant water use within achievable RC ranges. These can be generalized beyond application to a single species or substrate texture to inform the design and quality assurance of substrate placement for future landfill phytocaps.