Background In childhood acute lymphoblastic leukemia (ALL), radiotherapy for CNS prophylaxis is not used in frontline therapy anymore. Standard treatment for ALL nowadays consists of polychemotherapy. Therefore, assessment of potential chemotherapy‐induced cognitive side effects becomes important. Although neurotoxicity was demonstrated in cross‐sectional studies, longitudinal studies remain scarce. Procedure We evaluated intellectual development of 94 pediatric ALL patients between 1990 and 1997, diagnosed before the age of 12 years, treated according to the European Organisation for Research and Treatment of Cancer Children's Leukemia Group 58881 protocol. Three assessments of the Wechsler Intelligence Scale for Children Revised were performed since diagnosis, according to age. Using repeated measures regression analysis, we investigated the effect of gender (low versus increased) risk group, parents' education, age at diagnosis, intelligence quotient (IQ) subscale (verbal (VIQ) versus performance (PIQ) intelligence), and test session. Results PIQ scores were lower than VIQ at baseline (−5.3 points on average, p = 0.0032), yet PIQ increased more strongly (PIQ: +3.9 points per test session; VIQ: +0.8, p = 0.0079), so this baseline difference disappeared (p = 0.0079). There were no clear effects of gender (girls: +0.6 points; p = 0.78) or risk group (low risk: +1.5 points; p = 0.49), but IQ scores were higher when one parent had followed higher education (+9.5 points, p < 0.0001). Finally, diagnosis at younger age predicted lower IQ scores (−1.3 points per year, p = 0.0009). Conclusion Given that IQ scores did not decline, our findings demonstrate a stable pattern. However, the lower PIQ scores at baseline may indicate that performance functioning is vulnerable to acute neurotoxicity. Also, lower scores for younger patients highlight the stronger impact of the disease and/or treatment at younger age.Copyright © 2016 John Wiley & Sons, Ltd.