Examine psychosocial mediators of the effects of high vs. low-dose resistance training (RT) maintenance interventions among older (ages 50-69), overweight and pre-diabetic adults. Participants (N=123) completed a three-month supervised RT initiation phase and were subsequently randomised (time 1) to high or low-dose six-month unsupervised RT maintenance interventions (time 2), followed by a six-month no-contact phase (time 3). Online measures of putative mediators and RT behaviour. RT intervention condition (high vs. low dose) had significant effects on change from time 1 to time 2 in behavioural expectation, self-regulation and perceived satisfaction (f(2)=.04-.08), but not outcome expectancies, RT strategies or behavioural intentions (f(2)≤.02). Change in each of the putative mediators, except for outcome expectancies (f(2)≤.02), had significant effects on RT behaviour at times 2 (f(2)=.12-.27) and 3 (f(2)=.23-.40). In a multiple mediation model, behavioural expectation (f(2)=.11) and self-regulation (f(2)=.06) mediated the effects of RT intervention condition on time 2 RT behaviour, whereas perceived satisfaction did not (f(2)=.01). Self-regulation was a significant mediator of intervention effects on time 3 RT behaviour (f(2)=.11), but behavioural expectation and perceived satisfaction were not (f(2)=.04). Findings suggest that behavioural expectation and self-regulation are appropriate targets for RT maintenance interventions among at-risk older adults.