Resistance to acetyl-CoA carboxylase herbicides is documented in at least 43 grass weeds and is particularly problematic in Lolium, Alopecurus and Avena species. Genetic studies have shown that resistance generally evolves independently and can be conferred by target-site mutations at ACCase codon positions 1781, 1999, 2027, 2041, 2078, 2088 and 2096. The level of resistance depends on the herbicides, recommended field rates, weed species, plant growth stages, specific amino acid changes and the number of gene copies and mutant ACCase alleles. Non-target-site resistance, or in essence metabolic resistance, is prevalent, multigenic and favoured under low-dose selection. Metabolic resistance can be specific but also broad, affecting other modes of action. Some target-site and metabolic-resistant biotypes are characterised by a fitness penalty. However, the significance for resistance regression in the absence of ACCase herbicides is yet to be determined over a practical timeframe. More recently, a fitness benefit has been reported in some populations containing the I1781L mutation in terms of vegetative and reproductive outputs and delayed germination. Several DNA-based methods have been developed to detect known ACCase resistance mutations, unlike metabolic resistance, as the genes remain elusive to date. Therefore, confirmation of resistance is still carried out via whole-plant herbicide bioassays. A growing number of monocotyledonous crops have been engineered to resist ACCase herbicides, thus increasing the options for grass weed control. While the science of ACCase herbicide resistance has progressed significantly over the past 10 years, several avenues provided in the present review remain to be explored for a better understanding of resistance to this important mode of action.