Field experiments determined the critical period for weed control (CPWC) in grafted and nongrafted watermelon [Citrullus lanatus (Thumb.) Matsum. & Nakai] grown in plasticulture. Transplant types included ‘Exclamation’ seedless watermelon as the nongrafted control as well as Exclamation grafted onto two interspecific hybrid squash (ISH) rootstocks, ‘Carnivor’ and ‘Kazako’. To simulate weed emergence throughout the season, establishment treatments (EST) consisted of two seedlings each of common purslane (Portulaca oleracea L.), large crabgrass [Digitaria sanguinalis (L.) Scop.], and yellow nutsedge (Cyperus esculentus L.) transplanted in a 15 by 15 cm square centered on watermelon plants at 0, 2, 3, 4, and 6 wk after watermelon transplanting (WATr) and remained until the final watermelon harvest at 11 WATr. To simulate weed control at different times in the season, removal treatments (REM) consisted of two seedlings of the same weed species transplanted in a 15 by 15 cm square centered on watermelon plants on the same day of watermelon transplanting and allowed to remain until 2, 3, 4, 6, and 11 WATr, at which time they were removed. Season-long weedy and weed-free controls were included for both EST and REM studies in both years. For all transplant types, aboveground biomass of weeds decreased as weed establishment was delayed and increased as weed removal was delayed. The predicted CPWC for nongrafted Exclamation and Carnivor required only a single weed removal between 2.3 and 2.5 WATr and 1.9 and 2.6 WATr, respectively, while predicted CPWC for Kazako rootstock occurred from 0.3 to 2.6 WATr. Our study results suggest that weed control for this mixed population of weeds would be similar between nongrafted Exclamation and Exclamation grafted onto Carnivor. But the observed CPWC of Exclamation grafted onto Kazako suggests that CPWC may vary with specific rootstock–scion combinations.