Effects of no-till, minimum and conventional tillage practices and three weed control methods on weeds, growth, yield and yield components of upland rice were evaluated on a hydromorphic soil in the Guinea savanna, a low rainfall region of Nigeria. The weed control methods were: handweeding or herbicide application 15 and 40 days after planting and chemical control including preemergence treatment. Plant height and tillering at early growth stages of the crop were lower in no-till than in conventional and minimum tillage plots. Some level of soil manipulation seems to be necessary for satisfactory tillering in rice. The three tillage practices produced similar grain yield, weight of 100 grains, panicle number and number of filled grains panicle-1. Higher weed yields were recorded in no-till than in conventional and minimum tillage treatments. The herbicides used effectively controlled weeds under the three tillage practices and were comparable with two handweedings. Results indicate that herbicides hold promise for weed control in upland rice both in tilled and untilled soils.