We studied the effect of four weeding regimes (weed free, one manual weeding, one manual weeding + atrazine, and a weedy check) on larval density and leaf defoliation in four pear millet genotypes by the larvae of Oriental armyworm, Mythimna separata. Data were also recorded on the extent of larval parasitism under different weeding regimes, and the parasitoids involved. The leaf damage and larval densities were lower in weed free plots as compared to the weedy plots. This was also reflected in grain yield, as maximum grain yield was recorded in weed-free plots as compared to the weedy plots. Seven parasitoids (Cotesia ruficrus, Metopius rufus, Sturmiopsis inferens, Palexorista solemnis, P. laxa, Carcelia sp., and the entomopathogenic nematode Neoplectana sp. were recorded from M. separata larvae, of which M. rufus, Carcelia sp., and Neoplectana sp. were the most abundant. Parasitism by M. rufus was greater in plots with a weed cover and least in weed-free plots, while parasitsm by Carcelia sp. was lower in plots with one hand weeding than in weedy plots. Numerically, parasitism by Neopletana sp. was low in plots treated with atrazine, and maximum in plots weeded manually. Therefore, the minimum level of weeding, which does not affect the crop adversely should be undertaken to promote the biological control of M. separata in pearl millet.