A dataset comprised of road centerlines in Wyoming, USA, digitized to 2015 aerial photography from the National Agriculture Imagery Program. This dataset is an update to a former U.S. Geological Survey Data Series (“Large scale Wyoming transportation data: a resource planning tool”: O'Donnell and others, 2014) digitized to 2009 aerial photography. The U.S. Geological Survey Fort Collins Science Center created statewide roads data for the Bureau of Land Management Wyoming State Office using 2015 aerial photography from the National Agriculture Imagery Program. To ensure a systematic and repeatable approach of capturing roads on the landscape using on-screen digitizing from true color National Agriculture Imagery Program imagery, we developed a photogrammetry key and quality assurance/quality control protocols (O'Donnell and others, 2014). The updated statewide roads data will support the Bureau of Land Management’s resource management requirements with a standardized map product representing 2015 ground conditions. The data product, represented at 1:4,000 and +/- 10 meters spatial accuracy, contains 397,603 kilometers (O'Donnell and others, 2014 included 425,275 kilometers) within seven attribute classes. The quality control of these products indicated a 97.2 percent accuracy of aspatial information and 98.8 percent accuracy of spatial locations (an 86.3% accuracy was determined when considering spatial accuracy of mapped roads, commission errors, and omission errors). All errors identified during quality assurance assessments were corrected in the final product. Approximately 52 percent of the updated roads data were corrected for spatial errors of greater than 1 meters relative to the pre-existing road data (O'Donnell and others, 2014). Twenty percent of the updated roads involved correcting spatial errors of greater than 5 meters and 16 percent of the updated roads involved correcting spatial errors greater than 9 meters. The updated dataset includes the following improvements:1) Roads developed since 2009 but not roads developed after 2015; 2) Removal of commission errors (roads that were digitized but did not exist in imagery) was a principal focus of this dataset.