This systematic review with meta-analyses investigates the impact of resistance training (RT), using meta-regressions, on functional performance in frail and pre-frail adults aged ≥ 65 years to determine the key variables of RT. Ten randomized controlled trials involving 1303 participants were analyzed. Five studies assessed habitual walking speed (HWS), three studies evaluated performance in the timed-up-and-go test (TUG), three studies evaluated performance in the Short Physical Performance Battery (SPPB), and three studies assessed performance in the sit-to-stand test (STS). RT alone improved STS time and SPPB scores in frail and pre-frail older adults. RT improved STS performance (Effect Size (ES):− 0.536; 95% CI − 0.874 to − 0.199; p = .002) and led to a 2.261-point increase in SPPB performance (ES:1.682; 95% CI 0.579–2.786; p = .003). At least two weekly training sessions are required to increase SPPB scores, and three sessions seem to optimize the improvements. Higher training volume per exercise and volume per session reduce the gains in SPPB performance. We did not observe any association between different doses of RT and STS time improvements. RT alone positively influenced TUG performance only in community-dwelling older frail and pre-frail adults but not in institutionalized older individuals. RT alone did not improve the HWS compared to the non-active control group. •Progressive resistance training alone can improve sit-to-stand time and SPPB.•At least two weekly training sessions of RT are needed to increase SPPB scores•The effect of different doses of RT to promote STS improvements is still unclear.•High RT volume per session and per exercise may not be ideal for SPPB changes.•Resistance training alone did not promote differences in walking speed.