This study assessed male (n = 9) and female (n = 3) powerlifters' (18–49 years) ability to select loads using the repetitions in reserve–based rating of perceived exertion (RPE) scale for a single set for squat, bench press, and deadlift. Subjects trained 3× per week. For 3 weeks on nonconsecutive days in the weekly order of hypertrophy (8 repetitions at 8 RPE), power (2 repetitions at 8 RPE), and strength (3 repetitions at 9 RPE), using subject-selected loads intended to match the target RPE. Bench press and squat were performed every session and deadlift during strength and power only. Mean absolute RPE differences (|reported RPE—target RPE|) ranged from 0.22–0.44, with a mean of 0.33 ± 0.28 RPE. There were no significant RPE differences within lifts between sessions for squat or deadlift. However, bench press was closer to the target RPE for strength (0.15 ± 0.42 RPE) vs. power (−0.21 ± 0.35 RPE, p = 0.05). There were no significant differences within session between lifts for power and strength. However, bench press was closer (0.14 ± 0.44 RPE) to the target RPE than squat (−0.19 ± 0.21 RPE) during hypertrophy (p = 0.02). Squat power was closer to the target RPE in week 3 (0.08 ± 0.29 RPE) vs. 1 (−0.46 ± 0.69 RPE, p = 0.03). It seems that powerlifters can accurately select loads to reach a prescribed RPE. However, accuracy for 8-repetition sets at 8 RPE may be better for bench press compared with squat. Rating squat power-type training may take 3 weeks to reach peak accuracy. Finally, bench press RPE accuracy seems better closer rather than further from failure (i.e., 3-repetition 9 RPE sets vs. 2-repetition 8 RPE sets).