<b><i>Introduction:</i></b> Early identification of preterm children at high risk of intellectual disability (intelligence quotient [IQ] <70) or borderline intelligence (IQ = 70–84) is critical for different early intervention. We investigated whether early-life mental trajectories predict intellectual disability and borderline intelligence, respectively, among school-age preterm children. <b><i>Methods:</i></b> A multicenter study recruited preterm infants born at <32 weeks’ gestation between 2001 and 2014 in Taiwan who underwent mental assessments (Bayley Scales of Infant Development) at corrected ages 6, 12, and 24 months and IQs at age 5.5 years. Mental trajectories from ages 6 to 24 months determined using group-based trajectory modeling were employed to predict intellectual disability and borderline intelligence, respectively. Model fit was assessed using the area under the receiver operating characteristic curve (AUROC) and Akaike information criterion (AIC). <b><i>Results:</i></b> Among the 1,680 children enrolled, three mental trajectories were identified: high-stable (59.7%), high-declining (35.3%), and low-declining (5.0%), in which the borderline-intelligence/intellectual-disability rate was 14.1%/1.5%, 36.1%/13.7%, and 10.7%/82.1%, respectively. Compared with children with normal intelligence, the low-declining trajectory had 37.7-fold higher odds (95% confidence interval [CI], 26.3–48.1) for intellectual disability, and the high-declining trajectory had 4.4-fold higher odds (95% CI, 3.1–6.1) for borderline intelligence. Compared to the models with risk factors alone (AIC 1,791.2), the models that included both risk factors and trajectory groups had better overall performance (AIC 1,419.8) and increased prediction power for intelligence outcomes: low-declining trajectory for intellectual disability (AUROC increased from 0.81 to 0.92) and high-declining trajectory for borderline intelligence (AUROC increased from 0.68 to 0.75). <b><i>Conclusions:</i></b> Early-life mental trajectories help identify preterm children at risk of intellectual disability and borderline intelligence, respectively, at school age for timely intervention.