We developed a surface-associated activity trap (SAT) for sampling aquatic invertebrates in wetlands. We compared performance of this trap with that of a conventional activity trap (AT) based on non-detection rates and relative abundance estimates for 13 taxa of common wetland invertebrates and for taxon richness using data from experiments in constructed wetlands. Taxon-specific non-detection rates for ATs generally exceeded those of SATs, and largest improvements using SATs were for Chironomidae and Gastropoda. SATs were efficient at capturing cladocera, Chironomidae, Gastropoda, total Crustacea, and multiple taxa (taxon richness) but were only slightly better than ATs at capturing Dytiscidae. Temporal differences in capture rates were observed only for cladocera, Chironomidae, Dytiscidae, and total Crustacea, with capture efficiencies of SATs usually decreasing from mid-June through mid-July for these taxa. We believe that SATs may be useful for characterizing wetland invertebrate communities and for developing improved measures of prey available to foraging waterfowl and other aquatic birds.