Direct seeding methods to revegetate abandoned farmland were tested at a desert site west of Phoenix, Arizona. Native seeds were broadcast onto plots prepared by mulching, imprinting, chiseling, and fertilizing with phosphorous in a split-plot design. Each main plot was split into subplots that were not irrigated, irrigated with saline (3.25 dS/m) well water, or irrigated and hand weeded of Salsola iberica. Native seeds germinated poorly on all treatments, and three annual exotic weeds (Brassica nigra, S. iberica, and Schismus spp.) dominated the plots. None of the main plot treatments (mulching, imprinting, chiseling, or fertilizing) had a significant effect on seed germination or canopy cover. Irrigation increased plant cover on plots, but weeds dominated the cover (<4% native species, up to 50% weeds). Near the end of the second growing season a seed bank study was conducted in the greenhouse. Undisturbed desert soil had relatively few weed seeds and more native plant seeds than the disturbed agricultural soil samples, which had few viable native seeds and were dominated by Schismus spp., B. nigra, and S. iberica. The results illustrate the difficulty of establishing native plants on abandoned desert farmland due to the dominance of weedy species, the presence of salts in the soil, and the lack of adequate soil moisture in the treatments without supplemental irrigation.