To date, there is a lack of information about the optimal conditions of the warm-up to lead to a better performance in elite tennis players. The aim of this study was to compare the effects of two different warm-up protocols (dynamic vs. self-myofascial release with foam rolling) on neuromuscular variables associated with physical determinants of tennis performance. Using a crossover randomised experimental design, eleven professional men tennis players (20.6±3.5 years) performed either a dynamic warm-up (DWU) or a self-myofascial release with foam rolling (SMFR) protocol. DWU consisted of 8 min of dynamic exercises at increasing intensity and SMFR consisted of 8 min of rolling on each lower extremity unilaterally. Just before (baseline) and after completing warm-up protocols, players performed a countermovement jump (CMJ), the 5-0-5 agility test, a 10-m sprint test and the Straight Leg Raise and Thomas tests to assess range of motion. Compared to baseline, the DWU was more effective to reduce the time in the 5-0-5 test than SMFR (-2.23 vs. 0.44%, respectively, p= 0.042, ηp 2= 0.19). However, both warm-up protocols similarly affected CMJ (2.32 vs. 0.61%, p= 0.373, ηp 2= 0.04) and 10-m sprint time changes (-1.26 vs. 1.03%, p= 0.124, ηp 2= 0.11). Changes in range of motion tests were also similar with both protocols (p= 0.448–1.000, ηp 2= 0.00–0.02). Overall, both DWU and SMFR were effective to prepare well-trained tennis players for highly demanding neuromuscular actions. However, DWU offered a better preparation for performing change of direction and sprint actions, and hence, in high-performance tennis players, the warm-up should include dynamic exercises.