The aim of the study executed with the split-block method with 4 replications and conducted at the Experimental Station BASF Ltd. In Pągów in the years 2009–2011 was to evaluate the effect of different doses of herbicides on the reduction of weed, grain yield and structure of two varieties of winter wheat depending on the time of sowing. The studies tested two morphologically different varieties, Nutka and Ludwig. It has been shown that weed abundance on the control plots depends on both the sowing date as well as the cultivated variety of wheat. In the area where wheat was sown at the optimum sowing time 205.9 weeds per m2 occurred on plots with Ludwig variety, and 246.9 per m2 with Nutka while in the delayed sowing, 134.9 per m2 and 158.5 per m2, respectively. The effectiveness of the full doses of herbicides was much higher than of these reduced by 50%, which demonstrates the high sensitivity of weed species occurring in the study for active substances of herbicides. Based on the conducted survey it was found that both varieties of wheat reacted differently to grain yield on the control objects from the ones sprayed with herbicides. On the control plots, wheat sown at the optimal period produced smaller yields than wheat from the delayed sowing, while on the plots protected with herbicides the relations were inverse. In both sowing dates within the combination sprayed with herbicides no differences were observed in yielding wheat. It has been shown, however, that the variety Ludwig had a by 5.8%, and the variety Nutka by 16.7% higher yield from the timely sown crops than from the delayed sown. Analysis of the structure of yield has shown that grain yield depends in the highest degree on plant density per m2, and to a lesser extent on the weight of 1,000 grains, whereas the number of grains per spike remained unchanged.