Tillage regimes can influence weed population dynamics and, consequently, the choice of appropriate weed management practices. Studies were conducted in 2016 and 2017 in a long-term (36-yr) grain sorghum [Sorghum bicolor (L.) Moench ssp. bicolor] experiment at Texas A&M University, College Station, to determine the impact of long-term no-till (NT) and conventional till (CT) systems on weed species dynamics. Higher densities of johnsongrass [Sorghum halepense (L.) Pers.], prostrate spurge [Chamaesyce humistrata (Engelm. ex A. Gray) Small], waterhemp [Amaranthus tuberculatus (Moq.) Sauer], and henbit (Lamium amplexicaule L.) were recorded in the NT system compared with the CT system. Further, the NT system showed greater weed diversity (Shannon-Wiener index, H = 0.8) and species richness (S = 6.2), compared with the CT system (H = 0.6, S = 4.2). Seedling emergence of some dominant weed species was also delayed in the NT system. In the CT system, 50% emergence of S. halepense (8.5 C base temperature) and waterhemp (10 C base temperature) occurred at 59 and 63 growing degree days (GDD), respectively, whereas 68 and 75 GDD, respectively, were required in the NT system. Further, a greater proportion (61%) of the viable seedbank was present at the top 5 cm of the soil in the NT system compared with the CT system (46%). Overall, findings from this 36-yr-long tillage experiment have revealed that the NT system had greater weed densities (especially of the perennial weed S. halepense) and a high proportion of weed seeds (particularly small-seeded annuals) on the topsoil layer, corroborating some earlier reports that were based on short-term investigations. Findings indicate that growers transitioning to NT systems should be mindful of potential shifts in weed species dominance and develop appropriate management tactics.