This study evaluated the efficacy of a cognitive intervention for attentional control in older adults with mild cognitive impairment (MCI) with an executive deficit. It also sought to verify if the benefits of training generalised to primary and secondary outcome measures. Participants (n = 24) were randomly assigned to a training programme or active control condition. The experimental group completed a computer-based training programme involving Variable Priority (VP) coordination of both components of a dual task, to which was added a self-regulatory strategy designed to augment meta-cognition. The active control group performed Fixed Priority (FP) training: rote practice of the same dual task involving a visual detection task combined with an alpha-arithmetic task. Six one-hour training sessions were held three times a week for two weeks. Participants were tested pre- and post-training to detect improvement and transfer effects. Both groups improved on the visual detection and alpha-arithmetic tasks completed in focused attention, but only participants receiving VP training significantly improved their dual-task cost in accuracy for the visual detection task. As for transfer effects, both FP and VP training produced improvements on select outcome measures: focused attention, speed of processing, and switching abilities. No reliable advantage for generalisability of VP over FP training was found. Overall, these findings indicate that cognitive intervention may improve attentional control in persons with MCI and an executive deficit.