This study examined (a) the associations between family hostility (husband–wife marital hostility and child hostility) and middle-aged husbands’ and wives’ depressive symptoms over an 11-year time period and (b) the moderating influence of couples’ marital integration on these associations as measured by their joint activity. Higher order family-level latent constructs captured chronic husband–wife (marital) hostility using husbands’ and wives’ reports of chronic hostile interactions from 1990 to 1992, while a higher order latent construct of chronic child hostility toward parents was measured using parental reports of children’s hostile behaviors from 1990 to 1992. Structural equation modeling with data from 370 families depicted the longitudinal impact of family hostility on depressive symptoms of both husbands and wives in 2001 after accounting for earlier levels of depressive symptoms in 1991. Separate models were fit for couples with high and low levels of marital integration. For couples who experienced low levels of marital integration, chronic marital hostility and child hostility were related to depressive symptoms in husbands and wives. However, for those with high marital integration, these influences were largely diminished.