Field experiments were carried out at Krishi Nagar Research Farm, J.N. Krishi Vishwa Vidayalya, Jabalpur, (M.P.) during 2007-08 and 2008-09 to study weed composition and weed seed bank as influenced by tillage and crop establishment techniques in rice–wheat system. Sixteen treatments consisted with 4 tillage and planting management for both crop components under rice-wheat system were tested in strip plot design with 3 replications. Tillage and sowing methods were P1- direct drilling in dry field, P2- direct seeding of sprouted seeds through drum seeder in puddled field, P3- manual transplanting and P4- transplanting through self propelled transplanter (SPT) for rice cultivar ‘Kranti’ and T1- conventional tillage sowing, T2- zero till sowing, T3- strip till sowing and T4- bed planting for wheat cultivar ‘GW-273’. The total weed density and weed biomass at 30 DAS and maturity stages were significantly greater under direct drilling in dry field (DSR-P1) than other 3 sowing/planting methods of rice under puddled conditions (P2-direct seeding of sprouted seeds through drum seeder in puddled field, P3-manual transplanting and P4-transplanting through self propelled transplanter). The DSR-P1 had also higher weed seed counts on top layer of soil than other 3 tillage and sowing methods of rice. In wheat, intensity of grasses, sedges and other minor weeds was enhanced at maturity over their intensity at 30 DAS under conventional till sown wheat, while intensity of broad-leaved weeds (BLWs) declined at maturity over their intensity at 30 DAS. The higher weed seed count (40.9/m²) at top layer of soil was obtained extensively under zero-till sowing of wheat than conventional till sowing, strip till sowing and bed planting.