The critical period for weed control (CPWC) adds value to integrated weed management by identifying the period during which weeds need to be controlled to avoid yield losses exceeding a defined threshold. However, the traditional application of the CPWC does not identify the timing of control needed for weeds that emerge late in the critical period. In this study, CPWC models were developed from field data in high-yielding cotton crops during three summer seasons from 2005 to 2008, using the mimic weed, common sunflower, at densities of two to 20 plants per square meter. Common sunflower plants were introduced at up to 450 growing degree days (GDD) after crop planting and removed at successive 200 GDD intervals after introduction. The CPWC models were described using extended Gompertz and logistic functions that included weed density, time of weed introduction, and time of weed removal (logistic function only) in the relationships. The resulting models defined the CPWC for late-emerging weeds, identifying a period after weed emergence before weed control was required to prevent yield loss exceeding the yield-loss threshold. When weeds emerged in sufficient numbers toward the end of the critical period, the model predicted that crop yield loss resulting from competition by these weeds would not exceed the yield-loss threshold until well after the end of the CPWC. These findings support the traditional practice of ensuring weeds are controlled before crop canopy closure, with later weed control inputs used as required.