Monoculture crop production and prevailing farming practices have greatly reduced perennial plants on the landscape and nearly eliminated native Iowa prairie vegetation. The STRIPs (Science-based Trials of Row crops Integrated with Prairies) project is a watershed-scale experiment at the Neal Smith National Wildlife Refuge, in Jasper County, Iowa, US, in which strips of prairie vegetation were planted within watersheds of corn (Zea mays) and soybean (Glycine max) production to aid in soil and water conservation. The project includes 12 0.5- to 3.2-ha watersheds. Nine watersheds included buffer strips in one of three design treatments that varied the number and position of strips and/or the proportion of the watershed converted to buffer and three watersheds were 100% crop. The present study investigated: (1) If the design of prairie buffer strips influenced their vegetation; (2) If the vegetation of prairie buffer strips shifted over time; (3) If prairie buffer strips caused a weed problem in adjacent crop fields. From 2008-2011, the identity and percent cover of plant species within the buffer strips were surveyed, and from 2009-2011, the identity and percent cover of weed species within the cropped areas of the watersheds were surveyed. Differences among treatments and among years in plant species diversity, percent cover, and composition were analyzed using ANOVA and NMS. The design of buffer strips did not influence plant species diversity or composition; however, buffer strip vegetation did shift over time. In 2008, the strips had 38 species (in 6 m2) with 37% of the total plant cover composed of perennial species and 22% composed of native perennial species. By 2011, the strips had 55 species (in 6 m2) with 90% of the total plant cover composed of perennial species and 58% composed of native perennial species. In addition, NMS analyses indicated that the buffer strip plant community shifted from annual to perennial species. Within the crop, weed species richness and percent cover did not differ among watershed treatments, regardless of whether watersheds contained buffer strips or not. Prairie buffer strips greatly increased plant diversity in the watersheds; 380% more species were found in 6 m2 of prairie buffer than in 6 m2 of cropland. Within four years of establishment, the buffer strip vegetation was predominantly perennial and native species, the target vegetation for both ecohydrological functions (i.e., erosion control) and for conservation. Furthermore, weed species richness or prevalence did not differ between watersheds that incorporated prairie buffer strips versus 100% crop watersheds. Therefore, converting 10-20% of arable cropland to prairie buffer strips successfully reintroduced perennial species and conserved native Iowa prairie without causing a weed problem in adjacent crops.