The effects of different tillage systems and dose reduction on the efficacy of triasulfuron & chlortoluron mixture in the post-emergence control of annual broad-leaved weeds in winter wheat and spring barley were studied on lessive pseudogley soil in north-eastern Croatiaduring 1997–2000. Total dry biomass production in untreated plots was significantly influenced by tillage and it was lowest in continuous mouldboard ploughing (99 kg/ha), medium and similar in mouldboard ploughing/disk harrowing alternating every second year and in chisel ploughing (218 kg/ha) whereas the biomass was highest in continuous disk harrowing (422 kg/ha). Thereby the proportion of annual broad-leaved weed biomass was 70, 63, and 28%, respectively. Chenopodium album L., Ambrosia artemisiifolia L., Ch. polyspermum L. and Polygonum aviculare L. are the most abundant annual weed species in all tillage treatments. One half and one quarter of the recommended rate decreased the control efficacy of total weed biomass by 12 and 19%, respectively in wheat and by 6 and 15%, respectively in barley compared to the highest dose but they still provided a very good biomass control of main annual weeds (94–96 percentage units). The efficacy of reduced herbicide doses in the control of annual broad-leaved weeds did not vary significantly between tillage treatments and growing seasons. Significant interaction with continuous disk harrowing tillage and one-quarter herbicide dose was detected in the last year of wheat trial when perennial weeds increased their biomass proportion 8 times compared to four years before.