This study compared the effects of different sprint ball-based training programs on fitness-related performances in youth soccer players. Forty male players (age: 15.2 ± 0.6 yrs) participated in a short-term (8 weeks) randomized parallel fully controlled training study, with pre-to-post measurements. Players were randomly assigned to 3 sprint ball-based training groups: training with combined sprint (performing linear and change of direction sprints; CST), or using linear sprint (LST), or performing sprint with change of direction (CODT) and to a control group (CONT). Pre- and post-training players completed a test battery involving linear sprinting (10 and 20 m, and flying 10 m), 505 test (COD), 15 m test with ball (CODB), countermovement jump test (CMJ test) and maximal oxygen consumption (VO2max). All physical performance’ variables improved after the training intervention (all p < 0.05; ES ≥ 0.2). No main effect of groups was observed in linear sprinting, CMJ and VO2max (p > 0.05; ES < 0.2). A training group main effect was found (p < 0.0001; ES = 0.50) for COD with CODT induced better performance than LST and CONT (all p <0.0001; ES > 0.8). Also, a training group main effect (p = 0.009; ES = 0.35) was found for CODB with CODT elicited better performance than LST and CONT (all p = 0.001; ES > 0.80). Our findings suggest that ball-based training programs were more effective to improve fitness levels in youth players during the in-season period and that CODT modality was more effective to improve COD and CODB performances.