A trait‐based community assembly framework has great potential to direct ecological restoration, but uncertainty over how traits and environmental factors interact to influence community composition over time limits the widespread application of this approach. In this study, we examined how the composition of seed mixes and environment (north‐ vs. south‐facing slope aspect) influence functional composition and native plant cover over time in restored grassland and shrubland communities. Variation in native cover over 4 years was primarily driven by species mix, slope aspect, and a species mix by year interaction rather than an interaction between species mix and slope aspect as predicted. Although native cover was higher on wetter, north‐facing slopes for most of the study, south‐facing slopes achieved a similar cover (65%–70%) by year 4. While community‐weighted mean (CWM) values generally became more resource conservative over time, we found shifts in particular traits across community types and habitats. For example, CWM for specific leaf area increased over time in grassland mixes. Belowground, CWM for root mass fraction increased while CWM for specific root length decreased across all seed mixes. Multivariate functional dispersion remained high in shrub‐containing mixes throughout the study, which could enhance invasion resistance and recovery following disturbance. Functional diversity and species richness were initially higher in drier, south‐facing slopes compared to north‐facing slopes, but these metrics were similar across north‐ and south‐facing slopes by the end of the 4‐year study. Our finding that different combinations of traits were favored in south‐ and north‐facing slopes and over time demonstrates that trait‐based approaches can be used to identify good restoration candidate species and, ultimately, enhance native plant cover across community types and microhabitat. Changing the composition of planting mixes based on traits could be a useful strategy for restoration practitioners to match species to specific environmental conditions and may be more informative than using seed mixes based on growth form, as species within functional groups can vary tremendously in leaf and root traits.