Background Separation anxiety disorder is the most prevalent childhood anxiety condition, but no study assessed children for separation anxiety at preschool age and followed them longitudinally and directly until mid‐childhood/early adolescence. Methods Multi‐informant (children, teachers, family), multipoint (at age 8, 10, 12, 13) assessments of 1,290 children of the Quebec Longitudinal Study of Child Development, who had been categorized between age 1.5 and 6 into four specific separation anxiety trajectories (1, low‐persistent; 2, low‐increasing; 3, high‐decreasing, and the less common: 4, high‐increasing) by growth mixture modeling. Participants in the high‐increasing trajectory were compared to participants in the other three trajectories for: (a) child's internalizing and externalizing problem behavior; (b) physical health; (c) academic achievement; (d) maternal anxiety. Results Multivariate analyses of variance/covariance at separate time points showed the high‐increasing trajectory mostly associated with: (a) higher internalizing, but not externalizing, behavior; (b) worse academic achievement (most consistently by comparisons to the normative low‐persistent trajectory; (c) higher rates of maternal panic/agoraphobic anxiety; (d) worse physical health (most consistently by comparisons to the low‐persistent trajectory). The high‐increasing trajectory had twofold to threefold higher incidences of physical illnesses than the normative low‐persistent group; this was specific for headaches at age 12 years, chronic asthma at age 10 and 13, and having received asthma‐related medication during the past 12 months. Conclusions High‐increasing separation anxiety in preschool maintains longitudinal relationships to independent health and academic outcomes, at least until preadolescence. This knowledge can inform the deployment of clinical resources at the earlier signs of the more impairing manifestations.