Purpose: To provide a joint-level analysis of traditional (TS) and cluster (CS) set structure during the back-squat exercise. Methods: Eight men (24  y, 177.3 [7.9] cm, 82.7 [11.0] kg, 11.9 [3.5] % body fat, and 150.3 [23.0] kg 1-repetition maximum [1RM]) performed the back-squat exercise (80%1RM) using TS (4 × 6, 2-min interset rest) and CS (4 × [2 × 3], 30-s intraset rest, 90-s interset rest), randomly. Lower-limb kinematics were collected by motion capture, as well as kinetic data by bilateral force platforms. Results: CS attenuated the loss in mean power (TS −21.6% [3.9%]; CS −12.4% [7.5%]; P = .042), although no differences in gross movement pattern (sagittal-plane joint angles) within and between conditions were observed (P ≥ .05). However, joint power produced at the hip increased from repetition (REP) 1 through REP 6 during TS, while a decrease was noted at the knee. A similar pattern was observed in the CS condition but was limited to the hip. Joint power produced at the hip increased from REP 1 through REP 3 but returned to REP 1 values before a similar increase through REP 6, resulting in differences between conditions (REP 4, P = .018; REP 5, P = .022). Conclusions: Sagittal-plane joint angles did not change in either condition, although CS elicited greater power. Differing joint power contributions (hip and knee) suggest potential central mechanism that may contribute to enhanced power output during CS and warrant further study. Practitioners should consider incorporating CS into training to promote greater power adaptations and to mitigate fatigue.