Externalizing problems—like aggression and delinquency—become increasingly more stable and severe over time making early identification of risk factors essential for intervention and prevention efforts. Previous research has identified insecure parent–child attachments as a risk factor; however, additional research is necessary to understand not only how secure attachments buffer externalizing behavior, but also which insecure classifications (i.e., avoidant, ambivalent, or disorganized)—and to which parent—have the worst long-term prognoses. The current study utilized a six wave, multi-method, multi-reporter longitudinal study to explore the effect of early attachment relationships on externalizing trajectories of children and adolescents from 6- to 15-years-old. First, stability of mother–child and father–child attachments from Time 1 to Time 6 was assessed using a variety of attachment measures. Second, multilevel growth curve models were used to explore externalizing trajectories in two models, one with mother–child attachment classifications (secure, avoidant, ambivalent, or disorganized) and other with father–child attachment classifications as time-invariant covariates. Third, another multilevel growth curve model examined trajectories of externalizing symptoms with the number of secure attachments (i.e., two, one, or none) across both parents as a time-invariant covariate. Results showed that father–child attachment was moderately stable with no significant differences between assessments in the same time point or similar assessments over time; however, mother–child attachment stability appeared largely dependent upon the use of similar measures over time. Results of latent growth models showed that—whether the attachment relationship was with mother, father, or both parents—children did not differ on their initial externalizing levels or rates of change with two exceptions: children with avoidant or ambivalent mother–child attachments experienced a greater decrease in symptoms over time and when insecure classifications were collapsed into a single category, children with insecure attachments still experienced a faster decline in externalizing symptoms over time than did children with secure mother–child attachments. Results are discussed with regard to the importance of selecting similar attachment assessments in longitudinal investigations and the impact of divorce/separation on continuity of attachment.Finally, suggestions for future research are presented.