Compensatory wetland restoration is a critical component of holistic, ecosystem-level oil spill response strategies. An important goal of restoration is to rehabilitate food webs in impacted areas, but faunal assemblages and trophic dynamics are rarely included in post-restoration monitoring or assessments of success. Different approaches to wetland restoration, including variations in construction technique, may influence faunal recovery and trophic interactions. To explore these dynamics, we compared emergent plant communities, terrestrial arthropod assemblages, and trophic interactions in restored emergent marshes that were constructed in terrace and mound configurations and in reference areas in the Lower Neches Wildlife Management Area (Texas, United States). Plant community composition differed among all habitat types, with higher diversity on terraces and in reference marshes. Terrestrial arthropod abundance was similar among habitat types, but species composition was distinctly different among habitat types, especially at the herbivore level, where four of the eight herbivore species were found in either reference or restored sites, but not in both habitat types. Herbivores (primarily beetles and planthoppers) were more abundant than detritivores (midges and flies) in all habitat types. Predator (web-building and hunting spiders) abundance and species identity were similar among habitat types. Based on stable isotope analysis, trophic relationships differed among mounds, terraces, and reference areas. Herbivore diets were more variable at the reference sites than in either restored habitat type, aligning with higher plant diversity in reference areas. In contrast, detritivore diets were more variable at restored sites, where they were likely consuming food sources such as benthic algae or sediment organic matter. Predator diets were primarily comprised of herbivores in reference areas and detritivores in restored habitats. Overall, the restoration approaches supported abundant terrestrial arthropod assemblages, but species composition was different. In addition, trophic relationships differed between restored and reference areas, in part due to unique plant species assemblages at restored and reference sites. These results suggest that ecosystem restoration strategies that introduce geomorphological heterogeneity and plant diversity are more likely to support a diverse array of species and functions.