Two field experimental trials were carried out in central Italy, in 2005 and 2006, on biomass sorghum [Sorghum bicolor (L.) Moench] in order to assess weed control efficacy and selectivity to the crop of some pre- and post-emergence herbicides applied at different doses and in different mixtures. All herbicides showed good selectivity to the crop, although postemergence treatments showed higher transitory phytotoxicity effects than pre-emergence treatments, especially when high temperatures occurred after treatments, decreasing the selectivity of leaf herbicides (i.e. MCPA, 2,4-D, bromoxynil and dicamba). Considering pre-emergence applications, terbuthylazine alone against broadleaves or in mixtures at low doses with s-metolachlor against mixed infestations (grasses + broadleaves), seemed to be the best options to obtain a good selectivity to the sorghum and a high weed control level. Aclonifen showed some limits in terms of weed spectrum and could be recommended only against simplified broadleaf weed infestations without the presence of less susceptible weeds, like Amaranthus retroflexus, Portulaca oleracea and Solanum nigrum. Propachlor seemed not to be advisable due to the low efficacy against all the major broadleaf warmseason weed species in the Mediterranean areas. Considering post-emergence applications, all treatments gave quite similar results in terms of weed control, although, the mixture of terbuthylazine + bromoxynil seemed to be the best option due to bromoxinil’s higher efficacy than other foliar herbicides, such as MCPA, 2,4-D and dicamba, which can increase the efficacy of terbuthylazine alone especially under dry weather conditions. There were no significant differences in sorghum biomass between herbicide treatments, although, the more selective pre-emergence treatments showed, on average, a higher biomass yield value than the less selective post-emergence treatments. For these reasons, biomass values seemed to be more related to herbicide selectivity than to herbicide efficacy, especially in cases of scarce competitiveness of weed flora.