We review the significance of two forms of sexual conflict (different evolutionary interests of the two sexes) for genetic differentiation of populations and the evolution of reproductive isolation. Conflicting selection on the alleles at a single locus can occur in males and females if the sexes have different optima for a trait, and there are pleiotropic genetic correlations between the sexes for it. There will then be selection for sex limitation and hence sexual dimorphism. This sex limitation could break down in hybrids and reduce their fitness. Pleiotropic genetic correlations between the sexes could also affect the likelihood of mating in interpopulation encounters. Conflict can also occur between (sex-limited) loci that determine behaviour in males and those that determine behaviour in females. Reproductive isolation may occur by rapid coevolution of male trait and female mating preference. This would tend to generate assortative mating on secondary contact, hence promoting speciation. Sexual conflict resulting from sensory exploitation, polyspermy and the cost of mating could result in high levels of interpopulation mating. If females evolve resistance to make pre- and postmating manipulation, males from one population could be more successful with females from the other, because females would have evolved resistance to their own (but not to the allopatric) males. Between-locus sexual conflict could also occur as a result of conflict between males and females of different populations over the production of unfit hybrids. We develop models which show that females are in general selected to resist such matings and males to persist, and this could have a bearing on both the initial level of interpopulation matings and the likelihood that reinforcement will occur. In effect, selection on males usually acts to promote gene flow and to restrict premating isolation, whereas selection on females usually acts in the reverse direction. We review theoretical models relevant to resolution of this conflict. The winning role depends on a balance between the 'value of winning' and 'power' (relating to contest or armament costs): the winning role is likely to correlate with high value of winning and low costs. Sperm-ovum (or sperm-female tract) conflicts (and their plant parallels) are likely to obey the same principles. Males may typically have higher values of winning, but it is difficult to quantify 'power', and females may often be able to resist mating more cheaply than males can force it. We tentatively predict that sexual conflict will typically result in a higher rate of speciation in 'female-win' clades, that females will be responsible for premating isolation through reinforcement, and that 'female-win' populations will be less genetically diverse.