Adoption in ancient Rome was primarily a legal procedure for transferring paternal power (potestas) to ensure succession in the male line within Roman patriarchal society. The Latin word adoptio refers broadly to "adoption", which was of two kinds: the transferral of potestas over a free person from one head of household to another; and adrogatio, when the adoptee had been acting sui iuris as a legal adult but assumed the status of unemancipated son for purposes of inheritance. Adoptio was a longstanding part of Roman family law pertaining to paternal responsibilities such as perpetuating the value of the family estate and ancestral rites (sacra), which were concerns of the Roman property-owning classes and cultural elite. During the Principate, adoption became a way to ensure imperial succession.

In contrast to modern adoption, Roman adoptio was neither designed nor intended to build emotionally satisfying families and support childrearing. Among all social classes, childless couples or those who wanted to expand the size of their families instead might foster children. Evidence is meager for the adoptio of young children for purposes other than securing a male heir, and probably would have been employed mostly by former slaves legitimating the status of their own children born into slavery or outside a legally valid marriage.

Roman women could own, inherit, and control property as citizens, and therefore could exercise prerogatives of the paterfamilias pertaining to ownership and inheritance. They played an increasingly significant role in succession and the inheritance of property from the 2nd century BC through the 2nd century AD, but as an instrument for transferring paternal potestas, adoption was mainly a male-gendered practice.

Formal adoption was practiced primarily for financial, social, and political purposes among the property-owning classes. Free working people for whom these interests were minimal had little need of the cumbersome legal procedure and instead fostered if they wished to rear children. For the Romans, kinship was "biologically based but not biologically determined", and procedures such as adoption and divorce gave them greater latitude to restructure their families than was allowed in Christian Europe. Cicero said that adoption was an accepted way to ensure the hereditas (transmission) of three aspects of Roman family continuity: the family name (nomen), wealth (pecunia), and religious rites (sacra). Adoption was appropriate for a man who had no legitimate children, but if there were already legitimate heirs, adoption risked diluting their inheritance and the social status that came with it. Romans tended to prefer small families of two or three children for this reason, though premodern rates of neonatal and childhood mortality, along with other factors, could be an unsought brake on family size that jeopardized the family line. In adopting an adult heir, the father "could see what he was getting".

Adoption was carried out by the male who was head of his family, the paterfamilias, and his adopting did not make his wife a mother. Nor was marriage required; an adult bachelor could adopt in order to pass along his family name and potestas, as could a citizen eunuch (Latin spado).

A close relative was preferred as the adoptee, and a paterfamilias might adopt a grandson, especially if the grandson's father was not in the line of succession. The grandson might be his daughter's son, or the pater might have removed the boy's father from succession by emancipating him. One common pattern in Roman adoption was for a woman's childless brother to adopt one of her sons. A brother or cousin on the father's side might relinquish potestas over a son to provide a childless man with an adoptive heir. A pater who had no sons might adopt his daughter's husband to strengthen family lineage, but to avoid technical incest, he would first need to emancipate his daughter so that she was no longer legally a part of the family – the adoption would otherwise create a brother-sister relationship that Roman law regarded as consanguines, the same as blood ties. Adoption of a stepson from the wife's previous marriage was another strategy, if the stepson had no children; after adoption, his offspring would enter the line as grandchildren of the adopting paterfamilias.

The adoptee did not have to be a relative. Romans placed a high value on the social bonds of friendship (amicitia), and a childless man might adopt a friend or friend's son. Fostering was preferred to adopting children of "low" birth or unknown parentage, and in Roman Egypt it was unlawful to adopt a male foundling. The paterfamilias generally transmitted his estate to an adoptee of his own rank, or the adoptee acquired the social rank of the adoptive family, with some exceptions.

Most often adoption would have been a lateral move or a modest boost to the adoptee's standing and wealth, but a freedman could also be adopted. A slave might even be simultaneously manumitted and adopted by his former master, who became both his patron (patronus) and his "father". The adoption of a freedman placed his property under the control of his new paterfamilias; it no longer belonged to him, but it would return to him along with the rest of his inheritance. The choice of a freedman for adoption may have been motivated most often by gaining access to his resources rather than securing lineage.