In young elite athletes the influence of a sensorimotor training (SMT = balance training) on strength, jump height and spinal reflex excitability was compared with adaptations induced by strength training (ST). Seventeen athletes were randomly assigned to either a SMT or a ST group. Before and after 6 weeks of training, maximal isometric strength (MVC) and rate of force development (RFD (max)) were determined. Changes in jump height and EMG activity were assessed during squat- (SJ), countermovement- (CMJ) and drop-jump (DJ). To evaluate neural adaptations, H-reflex recruitment was recorded at rest and during dynamic activation of the plantarflexors following stance perturbation. MVC was enhanced after ST but not influenced by SMT. RFD (max) was not affected by any training. Both SMT and ST significantly improved jump performance in SJ, CMJ, and DJ. Maximum H-reflex to maximum M-wave ratios (H (max)/M (max)-ratios) at rest remained unchanged. During stance perturbation, H (max)/M (max)-ratios were significantly reduced following SMT whereas ST augmented H (max)/M (max)-ratios (p < 0.05). In contrast to other studies, no changes in RFD were found. This may be explained by methodological and/or training specific differences. However, both SMT and ST improved jump performance in well trained young athletes but induced opposing adaptations of the H (max)/M (max)-ratio when measured during dynamic contractions. These adaptations were task-specific as indicated by the unchanged reflexes at rest. Decreased spinal excitability following SMT was interpreted as the attempt to improve movement control, whereas augmented excitability following ST accounts for the effort to enhance motoneuron output. Functionally, our results emphasise that SMT is not only beneficial for prevention and rehabilitation but also improves athletic performance.