This study compared the neuromuscular control strategies and force-time profiles of three upper-limb pushing tasks-the standard push-up (SP), plyometric push-up (PP), and standard squat push-up (SSP)-representing the endurance-, power-, and strength-oriented continuum of upper-limb pushing performance. Fifteen male rugby athletes performed SP, PP, and SSP on dual force plates while surface electromyography (EMG) (12 muscles) and vertical forces were concurrently recorded. EMG signals were processed and decomposed using nonnegative matrix factorization to extract muscle synergies. Synergy modules were evaluated using cosine similarity and paired-samples t tests. One-dimensional Statistical Parametric Mapping (SPM1D) was employed to compare synergy primitives and force-time profiles across tasks. Two synergies reconstructed all tasks (VAF > 0.95). Synergy modules differed among conditions (cosine similarity < 0.90), with task-specific changes mainly involving distal forearm, scapular-trunk, and elbow-extensor muscle weightings. SPM1D identified task-specific differences in synergy primitive 1 during 0%-12% and 48% - 71%, and in primitive 2 during 71% - 100%, with PP and SSP exhibiting higher late-phase activation. PP generated higher force than SP during 13% - 79%, while SSP produced higher force during 33% - 36% and 56% - 83%, but lower force early and near takeoff. Despite sharing two synergies, the three pushing tasks exhibited distinct synergy structures, activation timing, and force-time profiles. PP emphasized rapid early-to-mid phase propulsion, whereas SSP relied on sustained late-phase force. These findings demonstrate task-specific neuromechanical regulation and may help inform exercise selection for upper-limb strength and power development in trained athletic populations.