Several studies have found that individuals select partners who resemble their parents. The evidence for this effect seems stronger in relation to opposite-sex than same-sex parents, although the ultimate-level biological explanations put forward to explain these preferences do not seem to require that they need to be built on the appearance of the opposite-sex parent, rather than any other immediate family member. We set out to revisit this question, while assessing face preferences rather than partner choice. Face preferences might uncover more subtle effects than partner choice, as they can elucidate preferences in an unconstrained environment. We presented participants with faces manipulated to resemble their mother, father or self, but did not find that they selected these faces as more suitable for relationships than control faces. However, consistent with previous work, participants who reported less childhood rejection by their opposite-sex parent selected faces that resembled that parent significantly more frequently than control faces. Taken together with previous work, opposite-sex parental faces seem more important than same-sex parental faces in shaping partner preferences, and childhood relationships seem to modify potential attraction to parent-resembling faces. Despite some inconsistent findings, this effect has been detected across the different methodologies used to assess preferences.