Objective: The aim of this study was to compare the effects of training using loads from a repetition maximum value (%1RM) and rating of perceived exertion (RPE) in elderly women. Methods: Twenty-five elderly women (60–75 years old) were randomly assigned to a group that trained using loads determined by 1RM test (G%; n = 12) or to a group that trained using loads determined by RPE (GPE; n = 13). Elderly women trained for 12 weeks using five exercises performed with 2–3 sets of 8–15 repetitions. Loads progressed from 45% to 75% of 1RM (G%) and from 13 to 18 from Rating Perceived Exertion of Borg Scale (GPE). The outcome measures, 1RM and maximum repetitions (RMs with 70% 1RM), were assessed before, between and after training programs. Results: Increased 1RM value and RMs were observed in both groups (20–42%, p < 0.001 and 56–76%, p < 0.001, respectively, for %G; and 17–56%, p < 0.001 and 47–106%, p < 0.001, respectively, for GPE), without differences between them. Conclusions: Prescribing loads using the RPE and 1RM might be similarly effective for training elderly women in order to promote strength gains. As a practical application, RPE could be an additional method to determine training loads. In spite of the promising results of the present study, it is not possible to state that the use of RPE is effective in monitoring loads during sub maximal strength training in elderly and more research must be carried out to confirm it.