A long-term experiment was conducted on Inceptisol at Indira Gandhi Krishi Vishwavidyalaya, Raipur, Chhattisgarh during 2010 to 2015 in direct-seeded rice (DSR)-chickpea cropping system to study the effect of continuous and rotational use of weed management practices on weed shift and productivity. No remarkable weed shift was visualized due to continuous or rotational application of combination of herbicides or manual weeding or its integration. However, appearance of Celosia argentea in unweeded plot was noticed in sixth year of DSR mainly due to its aggressive growth habit and non-submergence of rice field during Kharif 2015. The appearance of Celosia argentea suppressed the Alternanthera triandra. The Celosia argentea produced 8430 seeds/plant as against 1564 seeds/plant by Alternanthera triandra. Significantly higher seed yield of rice was registered under two hand weedings followed by oxadiargyl 80 g/ha fb bispyribac 25 g/ha. Seed yield of chickpea was higher under conventional tillage and continuous application of pendimethalin 1000 g/ha. Studies on weed seed bank suggested that although, there was no effect of different weed management treatments on seed bank up to 15 cm soil depth neither in DSR nor in chickpea at initial stage, but in unweeded control plot, there was perceptible variation in number of weed seeds of different annual weed species. The dominant weeds species were Celosia argentea (37.7%), Echinochloa colona (19.6%), Ischaemum rugosum (14.7%) and Cyperus iria (9.8%) over initially dominant species of Alternanthera triandra.