A growth model originally developed for grassland but also adjusted to Brassica, was modified and used to estimate the growth of annual weeds in spring barley. Growth rates, calculated for weed stands at the time of herbicide treatment in 17 experimental fields, were used to divide weed plant status in terms of growth rate into four classes: (1) no or poor establishment; (2) estimated weed population with constant growth rate; (3) with increasing growth rate; and (4) with decreasing growth rate. The statistical analyses showed that a herbicide application performed at increasing growth rate gave the best effect for two tested herbicides. The conclusion is that environmental conditions that promote growth rate of the weed stand will also increase the herbicide effect. The results were based on data from two field experimental series performed at six sites in southern Sweden during 1991–1994. Two herbicides, dichlorprop-P/MCPA and tribenuron-methyl + wetting agent were applied at four sub-normal doses on three alternate occasions. Portable weather stations recorded data on a number of weather parameters during the growing season.