Establishing on-farm pollinator habitat can mitigate native pollinator and pollination ecosystem service declines, and federal programs are available to provide technical and financial assistance to farmers interested in habitat installation. Although sufficient seedbed preparation to limit weed competition is thought to be the most important step toward achieving good pollinator habitat establishment, preparation recommendations vary and studies have not investigated seedbed preparation techniques in the context of the U.S. Department of Agriculture (USDA) Farm Bill programs for pollinator conservation. To address this, we assessed the effects of two seedbed preparation methods, conventional inversion tillage and no-till with herbicide, on wildflower establishment and weed competition during the first year after planting. Experiments were conducted in Blacksburg, VA, and were replicated in 2015 and 2016. In addition, we tracked seedbed preparation methods and pollinator habitat establishment on seven farms located on the Eastern Shore of Virginia and Maryland in 2016. The wildflower mix consisted of nine species of forbs and two species of grasses: the forbs were black-eyed Susan (Rudbeckia hirta L.), partridge pea [Chamaecrista fasciculata (Michx.) Greene], plains coreopsis (Coreopsis tinctoria Nutt.), lanceleaf coreopsis (Coreopsis lanceolata L.), purple coneflower [Echinacea purpurea (L.) Moench], narrowleaf mountain mint (Pycnanthemum tenuifolium Schrad.), wild bergamot (Monarda fistulosa L.), Maximilian sunflower (Helianthus maximiliani Schrad.), and showy tickseed [Bidens aristosa (Michx.) Britton] or Indian blanket (Gaillardia pulchella Foug.); the grasses were splitbeard bluestem (Andropogon ternarius Michx.) and little bluestem [Schizachyrium scoparium (Michx.) Nash]. In the experiments, wildflower establishment was greater after no-till with herbicide than after tillage preparation (2015: P = 0.09; 2016: P = 0.002). Predominating weed species varied by study, with more common lambsquarters (Chenopodium album L.) and hairy galinsoga [Galinsoga ciliata (Raf.) S.F. Blake] growth after tillage treatments, and more hairy vetch (Vicia villosa Roth) growth after no-till preparation. On-farm pollinator habitat establishment on the Eastern Shore was highly variable, but more wildflower forbs established after tillage-plus-herbicide than tillage-only seedbed preparations (P = 0.01). Across sites, we found a large degree of site-specific variation in wildflower establishment and predominant weed species.