In the future, maintaining food security will require an increase in rice productivity with less land, labor, and water. Agronomic approaches to reducing the water and labor demands of rice-based cropping systems include crop diversification to include crops with higher water productivity, direct-seeded rice, and irrigation scheduling based on soil water potential or alternate wetting and drying cycles. However, weeds are a major constraint to increasing rice productivity in these systems. A two-year study was conducted to identify the changes in weed population and to evaluate the efficacy of herbicides and their effect on crop performance in a sprinkler-irrigated maize-rice cropping system. The dominance of weed species shifted from a grass weed Echinochloa colona to broadleaved weeds, such as Hedyotis corymbosa, Lindernia spp., and Murdannia nudiflora. At 41 days after sowing (DAS) and at tasseling stage, glyphosate application provided a 90-100% decrease in total weed biomass of maize. The combined effects of all the herbicides applied in rice provided almost 100% control in total weed biomass at 40 DAS. However, the herbicide schedule in the cropping system was ineffective in controlling M. nudiflora. The weed control treatments had no significant effect on the yield of maize [8.7 and 11.5 (herbicide-treated) and 8.4 and 12.0 (partially weedy) Mg/ha in 2012 and 2013, respectively], but rice grain yield was influenced significantly. Rice grain yield in herbicide-treated plots in 2012 and 2013 were more than 1.5 and 8.5 times greater than yield achieved in the weedy plots, respectively. A significant negative relationship was found between rice grain yield and weed biomass in both years. Future research in maize-rice cropping systems should focus on the integration of appropriate agronomic practices with herbicide rotation to ensure that weed management strategies are sustainable and effective, and to avoid the risk of herbicide resistance developing in weed populations.