<p>Muscle synergies provide a framework to study motor control in healthy and clinical populations through electromyographic recordings during motor tasks. However, their variability has been poorly investigated. This study aimed to assess inter-session variability of muscle synergies during elevation tasks performed with the dominant and non-dominant upper limbs in healthy subjects. Fifteen young healthy adults performed explosive arm elevation movements along the sagittal, scapular and adduction planes, starting with the arm alongside the body and reaching a 90° angle relative to the trunk. Electromyographic signals were recorded from 20 muscles (agonist and postural muscles of upper limbs, trunk and lower limbs). Data were collected at baseline (T0) and after 1&#xa0;day (T1) without any intervention. Four synergies (Upper limb acceleration synergy, Upper limb deceleration synergy, Postural adjustment synergy 1 and Postural adjustment synergy 2) were extracted per session using non-negative matrix factorization. Cosine similarity and non-parametric statistics were used to assess inter-session consistency. Results showed significant inter-session similarity of muscle synergies (0.86 for the dominant limb, 0.78 for the non-dominant limb), compared to random similarity (0.48 and 0.47, respectively). The dominant arm exhibited higher reproducibility than the non-dominant arm (<i>p</i> = 0.002). Upper limb acceleration synergies were the most stable, whereas Postural adjustment synergies (1 and 2) showed greater variability, especially in the non-dominant arm. Muscle synergies during upper limb elevation exhibit moderate-to-high repeatability in healthy subjects, but non-negligible variability was found, reflecting physiological differences and potential methodological factors. Quantifying this variability may support interpretation of motor control changes in longitudinal studies.</p>