Major developments in attachment research over the past 2 decades have introduced parental mentalization as a predictor of infant–parent attachment security. Parental mentalization is the degree to which parents show frequent, coherent, or appropriate appreciation of their infants’ internal states. The present study examined the triangular relations between parental mentalization, parental sensitivity, and attachment security. A total of 20 effect sizes (N = 974) on the relation between parental mentalization and attachment, 82 effect sizes (N = 6,664) on the relation between sensitivity and attachment, and 24 effect sizes (N = 2,029) on the relation between mentalization and sensitivity were subjected to multilevel meta-analyses. The results showed a pooled correlation of r = .30 between parental mentalization and infant attachment security, and rs of .25 for the correlations between sensitivity and attachment security, and between parental mentalization and sensitivity. A meta-analytic structural equation model was performed to examine the combined effects of mentalization and sensitivity as predictors of infant attachment. Together, the predictors explained 12% of the variance in attachment security. After controlling for the effect of sensitivity, the relation between parental mentalization and attachment remained, r = .24; the relation between sensitivity and attachment remained after controlling for parental mentalization, r = .19. Sensitivity also mediated the relation between parental mentalization and attachment security, r = .07, suggesting that mentalization exerts both direct and indirect influences on attachment security. The results imply that parental mentalization should be incorporated into existing models that map the predictors of infant–parent attachment.