Background Cognitive impairments, including delirium, are common after coronary artery bypass grafting (CABG). Improving cognition pre- and post-operatively using computerised cognitive training (CCT) may be an effective approach to improve cognitive outcomes in CABG patients. Objectives Investigate the effect of remotely supervised CCT on cognitive outcomes, including delirium, in older adults undergoing CABG surgery. Methods Thirty-six participants, were analysed in a single-blinded randomised controlled trial (CCT Intervention: n = 18, Control: n = 18). CCT was completed by the intervention group pre-operatively (every other day, 45–60-minute sessions until surgery) and post-operatively, beginning 1-month post-CABG (3 x 45–60-minute sessions/week for 12-weeks), while the control group maintained usual care plus weekly phone calls. Cognitive assessments were conducted pre- and post-operatively at multiple follow-ups (discharge, 4-months and 6-months). Post-operative delirium incidence was assessed daily until discharge. Cognitive change data were calculated at each follow-up for each cognitive test (Addenbrooke’s Cognitive Examination III and CANTAB; z-scored). Results Adherence to the CCT intervention (completion of three pre-operative or 66% of post-operative sessions) was achieved in 68% of pre-CABG and 59% of post-CABG participants. There were no statistically significant effects of CCT on any cognitive outcome, including delirium incidence. Conclusion Adherence to the CCT program was comparatively higher than previous feasibility studies, possibly due to the level of supervision and support provided (blend of face-to-face and home-based training, with support phone calls). Implementing CCT interventions both pre- and post-operatively is feasible in those undergoing CABG. No statistically significant benefits from the CCT interventions were identified for delirium or cognitive function post-CABG, likely due to the sample size available (study recruitment greatly impacted by COVID-19). It also may be the case that multimodal intervention would be more effective.