Understanding how athletes reduce motor variability in redundant tasks contributes to improving sports performance and elucidating human motor control mechanisms. This study aimed to clarify how experienced basketball players use two hierarchical redundancies, ball-level and body-level, for accurate free-throw shooting as no study has simultaneously examined how these hierarchical redundancies are used. Experienced basketball players (intermediate-level and top-level) participated in a free-throw shooting experiment using a motion capture system under two conditions: with feedback (FB) and no-feedback (NF) conditions. To quantify the coordination, the solution manifold and tolerance, noise, and covariation analysis were used for the ball-level redundancy, while uncontrolled manifold analysis and the covariation by randomization method were used for the body-level redundancy. The ball-level analysis revealed that the covariation and noise components were related to the performance, and that the noise component showed a larger contribution to performance than the covariation component, indicating that the reproducibility of the release parameters has a larger room for improvement than coordination. The coordination of release parameters was not significantly different between the FB and NF conditions, indicating that the effect of performance feedback on coordination is unclear. The body-level analysis revealed that interjoint coordination reduced the variability of the longitudinal positions of the hand and fingertip, showing that interjoint coordination improves the reproducibility of the ball-release position, especially in the direction that strongly affects the motor performance. In conclusion, interjoint coordination improved the reproducibility of ball-release parameters, which enhanced motor performance in basketball free-throw shooting.