This study examines the effects of two strength training (ST) programs, one based on mean propulsive velocity (MPV) and another under the traditional method, the % one-repetition maximum (1 RM), on neuromuscular performance and muscle composition in girls who play soccer. Fifty players with an average age of 13.6 +/- 1.2 years participated in the study and were randomly assigned into three groups: a maximal execution velocity training group (VG, n=15), a maximal strength group (RMG, n=13), and a control group (CG, n=18). The study was developed for a period of twelve weeks during regular team training to prepare for the season. The VG and RMG groups performed additional strength or muscle power training three times a week, including movements of full squat and pedaling on a cycle ergometer. The two types of training groups and the control group exhibited significant gains. However, the greatest increases were achieved with VG training, with significant increases (p < 0.000) in maximal strength, (p < 0.000) squat power, (p < 0.000) velocity over 30 m, (p < 0.000) cycle ergometer power, and (p < 0.008) lower limb muscle mass. Statistically significant differences were observed between VG and RMG in countermovement jump (CMJ) (p < 0.008) and squat power (p < 0.01) tests, between VG and CG in CMJ (p < 0.01), squat power (p < 0.000), and maximal squat strength (p < 0.000), and between RMG and CG in maximal squat strength (p < 0.000) only. These findings might indicate that high-velocity ST can be performed simultaneously with regular training to improve the explosive actions of soccer players.