A field experiment was performed to determine the effects of different times and frequencies of handweeding on grain yields in upland rice. Five weeding treatments were done. These treatments consisted of single weedings at 20 and 35 days after emergence (d.a.e.) and double weedings at 10 and 15 d.a.e., 20 and 35 d.a.e. and 30 and 45 d.a.e. Two controls were made: a weed‐free control where weeds were continuously removed and a weedy control where no handweeding was done. Aboveground destructive harvests on the controls were used to determine dry matter partitioning. These were performed at 10, 30, 50, 70 and 90 d.a.e. and at harvest. Rice and weed heights were measured for the controls and all treatments. Grain yields were compared and analyzed statistically. Absence of weed control produced only 50 kg/ha of grain. Single weeding increased yield to about 2000 kg/ ha and double weeding to about 2500 kg/ha. However, no significant differences were found between treatments of the same weeding frequency. Complete weed control produced 3000 kg/ha of grain. During grain formation, stem weight of weed‐free rice declined by 24% as compared to 74% in weedy rice.