Abstract Background In their recent paper, Del Giudice and Haltigan argue that attachment in childhood and attachment representations in adulthood are influenced by the cognitive capabilities of children and parents, that would causally link parents' attachment states of mind to children's attachment. In the current pre‐registered study, we empirically explored the idea of an association between attachment and cognition using phenotypic child IQ and parent and child IQ‐related polygenic scores as predictors of children's attachment behavior and attachment representations. Methods We used data from the Leiden Consortium on Individual Development study (L‐CID, n = 992), a two‐cohort longitudinal twin study, in which attachment representations were measured in parents and their 9‐year‐old children using the Attachment Script Assessment (ASA). Polygenic scores of IQ were computed for parents and their children using PRSice‐2 and phenotypic child IQ was measured as well. We split the twin sample in two groups randomly to prevent non‐independence of data and conducted structural equation models. Results Neither parental nor child polygenic scores of IQ predicted representations of attachment. In one cohort, phenotypically measured IQ predicted attachment. Conclusions This preliminary study did not find convincing support for a role of IQ in the intergenerational transmission of attachment.