To investigate the long-term efficacy of computerized cognitive training in improving cognitive outcomes among childhood cancer survivors. Methods Sixty-eight survivors of childhood acute lymphoblastic leukemia (ALL) or brain tumor (BT) were randomly assigned to computerized cognitive intervention (23 ALL/11 BT, age = 12.21 ± 2.47) or a waitlist control group (24 ALL/10 BT, age = 11.82 ± 2.42). Cognitive assessments were completed pre-, immediately post-, and 6 months postintervention. Results A prior report showed training led to immediate improvement in working memory, attention and processing speed. In the current study, piecewise linear mixed effects modeling revealed that working memory and processing speed were unchanged from immediate to 6 months postintervention (intervention β = −.04 to .01, p = .26 to .95; control β = −.06 to .01, p = .23-.97), but group differences on an attention measure did not persist. Conclusion Cognitive benefits are maintained 6 months following computerized cognitive training, adding to potential clinical utility of this intervention approach.