IntroductionNovel, scalable, low‐cost interventions are needed to reduce harmful drinking amongst middle–older adults. Approach bias modification (ApBM) is a promising form of cognitive training for preventing/reducing alcohol use that can be delivered via smartphone. This study explored the acceptability and preliminary effectiveness of smartphone delivered and personalised ApBM amongst Australians ≥55 years, an age cohort at risk of alcohol‐related harms.MethodsSecondary analyses in a middle–older adult subsample (≥55 years, n = 289) of an open‐label pilot study using a retrospective, repeated measures design. We explored acceptability (adherence, user mobile acceptability ratings, free‐text responses) and preliminary effectiveness (changes in drinking quantity and frequency, craving, dependence and proportion drinking within government‐recommended guidelines) of two sessions/week over 4 weeks of evidence‐based ApBM training, adapted to include personalisation and smartphone delivery amongst Australians ≥55 years.ResultsAlthough minor adaptations to training were suggested, the intervention was acceptable amongst survey completers, with 72% training adherence. Relative to baseline, there was a significant increase in the proportion of drinking within recommended single‐session and weekly guidelines post‐training (from 25% to 41% and 6% to 28%, respectively, p < 0.001), with past‐week standard drinks significantly decreasing by 18% (p < 0.001) and significant reductions in drinking days, mean craving and dependence scores (p < 0.001).Discussion and ConclusionsFindings suggest smartphone ApBM is acceptable amongst middle‐to‐older aged Australians and may support this ‘at risk’ cohort to remain within government‐recommended alcohol consumption guidelines to optimise healthy aging, although, in the context of a single‐arm study, preliminary results should be interpreted cautiously.