Background Until now, working memory training has not reached sufficient evidence as effective treatment for ADHD core symptoms in children with ADHD; for young children with ADHD, no studies are available. To this end, a triple‐blind, randomized, placebo‐controlled study was designed to assess the efficacy of Cogmed Working Memory Training (CWMT) in young children with ADHD. Methods Fifty‐one children (5–7 years) with a DSM‐IV‐TR diagnosis of ADHD (without current psychotropic medication) were randomly assigned to the active (adaptive) or placebo (nonadaptive) training condition for 25 sessions during 5 weeks. The compliance criterion (>20 sessions) was met for 47 children. The primary outcome measure concerned the core behavioural symptoms of ADHD, measured with the ADHD Rating Scale IV (ADHD‐RS). Secondary outcome measures were neurocognitive functioning, daily executive functioning, and global clinical functioning. The influence of the increase in difficulty level (Index‐Improvement) for the treatment group was also analysed. Clinical trial registration information – ‘Working Memory Training in Young ADHD Children’; www.clinicaltrials.gov; NCT00819611. Results A significant improvement in favour of the active condition was found on a verbal working memory task (p = .041; adapted Digit Span WISC‐III, backward condition). However, it did not survive correction for multiple testing. No significant treatment effect on any of the primary or other secondary outcome measurements was found. The Index‐Improvement significantly contributed to ADHD‐RS and the Behavior Rating Inventory of Executive Function, both rated by the teacher, but revealed no significant group difference. Conclusions This study failed to find robust evidence for benefits of CMWT over the placebo training on behavioural symptoms, neurocognitive, daily executive, and global clinical functioning in young children with ADHD.