This dataset contains vegetation data collected at a variety of watershed restoration sites across southeastern Arizona over 5 years. The semiarid habitats in the Madrean Archipelago Ecoregion, which extends from southern Arizona into northern Mexico, are facing many challenges from climate change to land use change which threaten the ecological and cultural values of the region. Watershed restoration practitioners use a variety of techniques such as gabions, check dams, and cross vanes to reduce the effects of these threats and improve or maintain watershed function. Since vegetation dynamics in the area are driven by water availability, these restoration techniques appear to have secondary effects on the vegetation of a watershed. To evaluate and quantify these effects, vegetation data was collected at 5 restoration sites across southeastern Arizona. Three of these sites, Barboot (BB), Silver Creek (SC), and Vaughn Canyon (VC), were restored just before sampling occurred allowing for an "After-Control-Impact" design to assess the change in vegetation. Data was collected at SC and VC for 5 years while only 4 years of data was collected at BB due to a lack of continued change. At these sites treatment plots were established at structures within stream channels while control plots were established at locations in reaches without structures within the same drainages. At Deep Dirt (DD), the project drainage was fairly short and there was no area within it or in a similar drainage that did not have restoration structures. This prevented the installation of control plots, instead data was collected at two different types of structures for 5 years. At Cienega Ranch (CR), restoration structures had not been installed when data was collected. Plots were distributed in the project reaches and in a nearby control reach. One year of baseline data has been collected at CR which will allow for a “Before-After-Control-Impact” study design. At all project sites, each plot was stratified in zones based on hydrological position relative to the restoration structure and proximity to the structure. Nested quadrats were used to quantify abundance as well as basal and foliar cover. Subplots were used to quantify species composition. Data was collected for perennial species and non-native species during the summer monsoon season from August – September, sometimes extending into early October. This vegetation data is provided in 5 CSV files which form a relational database which is explained in further detail in "RelationalDatabase_VegetationResponseWatershedRestoration.xml". "Relationships.pdf" is an illustration of the relationships in the database. "StudySampleDesign.pdf" contains figures illustrating plot design and nested quadrat design to clarify the data collection process. "mtbl_Plants.csv" is the master plant list that includes taxonomic and other information for Arizona species; the development of this dataset is explained in further detail in "mtbl_Plants.xml". Every plot and nested quadrat was photodocumented; photos can be found in the "Photos_VegetationResponseWatershedRestoration" directory and process explained in the associated metadata file "Photos_VegetationResponseWatershedRestoration.xml". This data release can be downloaded with or without the photos. "VegetationResponseWatershedRestoration.zip" includes the photos; "VegetationResponseWatershedRestoration_noPhotos.zip" does not include the photos for faster download times.