The aim of the present study was to compare the effects of 3 training protocols (plyometric [PLYO], agility [AG], or repeated shuttle sprints [RS]) on physical performance in the same population of young soccer players. Forty-two youth-level male players (13.6±0.3-years; 1.65±0.07 m; 54.1±6.5 kg; body fat: 12.8±2.6%) participated in a short-term (6-week) randomized parallel fully controlled training study (pre-to-post measurements): PLYO group, n=10; AG group, n=10; RS group, n=12; and control group [CON] n=10. PLYO training = 9 lower limb exercises (2-3 sets of 8-12 repetitions). The AG group performed planned AG drills and direction changes. RS training consisted of 2-4 sets of 5-6x 20 to 30 m shuttle sprints (20 seconds recovery in between). Progressive overload principles were incorporated into the programme by increasing the number of foot contacts and varying the complexity of the exercises. Pre/post-training tests were: bilateral standing horizontal jump, and unilateral horizontal jumps, sprint (30 m with 10 m lap time), agility (20 m zigzag), and repeated sprint ability (RSA) (i.e. 6x30 m shuttle sprints: 2x15 m with 180° turns). Significant main effects for time (i.e. training application) and group (training type) were detected. Improvements in horizontal jumping were higher (p<0.01: ES=large) in PLYO. The RS group improved significantly more (p<0.01; ES=large) than other groups: 30 m sprint, RSAbest and RSAmean performances. Significantly greater increases in 20 m zigzag performance were observed following AG and RS training (4.0 and 3.8%, respectively) compared with PLYO (2.0%) and CON training (0.8%). No significant differences were reported in the RSAdec between groups. Elite young male soccer players’ physical performances can be significantly and specifically improved either using PLYO or AG or RSA training over short-term in-season training.