Sexually reproducing animals, plants, fungi and protists are thought to have evolved from a common ancestor that was a single-celled eukaryotic species. Sexual reproduction is widespread in eukaryotes, though a few eukaryotic species have secondarily lost the ability to reproduce sexually, such as Bdelloidea, and some plants and animals routinely reproduce asexually (by apomixis and parthenogenesis) without entirely having lost sex. The evolution of sexual reproduction contains two related yet distinct themes: its origin and its maintenance. Bacteria and Archaea (prokaryotes) have processes that can transfer DNA from one cell to another (conjugation, transformation, and transduction), but it is unclear if these processes are evolutionarily related to sexual reproduction in Eukaryotes. In eukaryotes, true sexual reproduction by meiosis and cell fusion is thought to have arisen in the last eukaryotic common ancestor, possibly via several processes of varying success, and then to have persisted.

Since hypotheses for the origin of sex are difficult to verify experimentally (outside of evolutionary computation), most current work has focused on the persistence of sexual reproduction over evolutionary time. The maintenance of sexual reproduction (specifically, of its dioecious form) by natural selection in a highly competitive world has long been one of the major mysteries of biology, since both other known mechanisms of reproduction – asexual reproduction and hermaphroditism – possess apparent advantages over it. Asexual reproduction can proceed by budding, fission, or spore formation and does not involve the union of gametes, which accordingly results in a much faster rate of reproduction compared to sexual reproduction, where 50% of offspring are males and unable to produce offspring themselves. In hermaphroditic reproduction, each of the two parent organisms required for the formation of a zygote can provide either the male or the female gamete, which leads to advantages in both size and genetic variance of a population.

Sexual reproduction therefore must offer significant fitness advantages because, despite the two-fold cost of sex (see below), it dominates among multicellular forms of life, implying that the fitness of offspring produced by sexual processes outweighs the costs. Sexual reproduction derives from recombination, where parent genotypes are reorganised and shared with the offspring. This stands in contrast to single-parent asexual replication, where the offspring is always identical to the parents (barring mutation). Recombination supplies two fault-tolerance mechanisms at the molecular level: recombinational DNA repair (promoted during meiosis because homologous chromosomes pair at that time) and complementation (also known as heterosis, hybrid vigour or masking of mutations).

Reproduction, including modes of sexual reproduction, features in the writings of Aristotle; modern philosophical-scientific thinking on the problem dates from at least Erasmus Darwin (1731–1802) in the 18th century. August Weismann picked up the thread in 1885, arguing that sex serves to generate genetic variation, as detailed in the majority of the explanations below. On the other hand, Charles Darwin (1809–1882) concluded that the effect of hybrid vigor (complementation) "is amply sufficient to account for the ... genesis of the two sexes". This is consistent with the repair and complementation hypothesis, described below. Since the emergence of the modern evolutionary synthesis in the 20th century, numerous biologists including W. D. Hamilton, Alexey Kondrashov, George C. Williams, Harris Bernstein, Carol Bernstein, Michael M. Cox, Frederic A. Hopf and Richard E. Michod – have suggested competing explanations for how a vast array of different living species maintain sexual reproduction.

The concept of sex includes two fundamental phenomena: the sexual process (fusion of genetic information of two individuals) and sexual differentiation (separation of this information into two parts). Depending on the presence or absence of these phenomena, all of the existing forms of reproduction can be classified as asexual, hermaphrodite or dioecious. The sexual process and sexual differentiation are different phenomena, and, in essence, are diametrically opposed. The first creates (increases) diversity of genotypes, and the second decreases it by half.

Reproductive advantages of the asexual forms are in quantity of the progeny, and the advantages of the hermaphrodite forms are in maximal diversity. Transition from the hermaphrodite to dioecious state leads to a loss of at least half of the diversity. So, the primary challenge is to explain the advantages given by sexual differentiation, i.e. the benefits of two separate sexes compared to hermaphrodites rather than to explain benefits of sexual forms (hermaphrodite + dioecious) over asexual ones. It has already been understood that since sexual reproduction is not associated with any clear reproductive advantages over asexual reproduction, there should be some important advantages in evolution.^{[better source needed]}

For the advantage due to genetic variation, there are three possible reasons this might happen. First, sexual reproduction can combine the effects of two beneficial mutations in the same individual (i.e. sex aids in the spread of advantageous traits) without the mutations having to have occurred one after another in a single line of descendants.^{[unreliable source?]} Second, sex acts to bring together currently deleterious mutations to create severely unfit individuals that are then eliminated from the population (i.e. sex aids in the removal of deleterious genes). However, in organisms containing only one set of chromosomes, deleterious mutations would be eliminated immediately, and therefore removal of harmful mutations is an unlikely benefit for sexual reproduction. Lastly, sex creates new gene combinations that may be more fit than previously existing ones, or may simply lead to reduced competition among relatives.

For the advantage due to DNA repair, there is an immediate large benefit of removing DNA damage by recombinational DNA repair during meiosis (assuming the initial mutation rate is higher than optimal), since this removal allows greater survival of progeny with undamaged DNA. The advantage of complementation to each sexual partner is avoidance of the bad effects of their deleterious recessive genes in progeny by the masking effect of normal dominant genes contributed by the other partner.