Background Cognitive complaints are common amongst breast cancer survivors, and no standard treatment exists. The present study evaluates whether web‐based cognitive training can alleviate subjectively reported and objectively assessed cognitive complaints in a sample of breast cancer survivors. The primary and secondary outcomes were an objective measure of working memory and a measure of perceived cognitive functioning. Additional outcomes were neuropsychological tests of memory, executive function, working memory and questionnaire‐based assessment of anxiety, depression and somatization. Methods A total of 157 female breast cancer survivors were recruited from an existing cohort and through announcements in open access cancer‐related Internet fora and randomly allocated to either web‐based cognitive training (eCogT) with telephone support (n = 94) or a waitlist control (WLC) condition (n = 63). eCogT encompassed 30 training sessions over 6 weeks. Neuropsychological assessments were undertaken over the telephone, and questionnaire data was collected online. Data was collected at baseline, post‐intervention and at 5‐month follow‐up. Results Mixed linear models revealed no statistically significant change in primary or secondary outcome at follow‐up in either group. Statistically significant improvements (p 0.040–0.043) were found in the eCogT group for verbal learning and on a working memory test. Conclusions Web‐based cognitive training did not result in improvements of the primary or secondary outcome. Improved performance was observed on verbal learning and working memory. These effects were observed at 5‐month follow‐up, indicating long‐term effects of training. The intervention may be applied in a clinical setting at low cost and without risk of adverse effects.© 2016 The Authors Psycho‐Oncology Published by John Wiley & Sons Ltd.