Prenatal maternal stress (PNMS) has been associated with postnatal behavioral alterations that may be partly explained by interactions between the hypothalamic–pituitary–adrenal (HPA) and hypothalamic–pituitary–gonadal (HPG) axes. Yet it remains unclear whether PNMS leads to enduring HPA–HPG alterations in the offspring, and whether HPA–HPG interactions can impact behavior during development, in particular levels of aggression in childhood. Here we investigated the relationship between a marker for HPG axis function (baseline testosterone) and a marker for HPA axis response (cortisol area under the curve) in 11½-year-olds whose mothers were exposed to the 1998 Quebec ice storm during pregnancy (n = 59 children; 31 boys, 28 girls). We examined (a) whether the degree of objective or subjective PNMS regulates the testosterone–cortisol relationship at age 11½, and (b) whether this testosterone–cortisol relationship is associated with differences in aggressive behavior. We found that, at lower levels of subjective PNMS, baseline testosterone and cortisol reactivity were positively correlated; in contrast, there was no relationship between these hormones at higher levels of subjective PNMS. Cortisol response moderated the relationship between testosterone and aggression. These results support the notion PNMS may explain variance in fetal HPA–HPG interactions, and that these interactions may be associated with aggressive behavior in late childhood.