Soccer-specific training is easily associable to players’ sprint abilities demonstrated during a match. However, no clear evidence has been provided to show whether this approach is more effective than training focused on running techniques for sprints in prepubescent soccer players. Thus, the present study aimed at comparing the effects of these two training approaches on prepubescent soccer players’ sprint performances. Ninety-five players (10±2 years) competing in local (Piedmont, Italy) Under-9 (N=21), -10 (N=24), -11 (N=25) and -13 (N=25) championships were recruited for the study. Sixty-three and 32 players were included in the running training group (RTG) and soccer-specific group (SSG), respectively. Before (PRE) and after (POST) the training period (2 weekly sessions for 12 weeks), sprint abilities were evaluated by means of four 20-m sprint tests: linear sprint (20-mL), linear sprint with ball possession (20-mLB), sprint with change of direction (20- mCoD), sprint with change of direction and with ball possession (20-mCoDB). A linear mixed model was applied to evaluate differences (P≤0.05) between the RTG and SSG in the four tests and categories, comparing PRE and POST performances. A main effect emerged for the interaction between groups, sessions (p=0.014; Between PRE ES range=0.03, 0.85; Within PRE-POST ES range=-0.45, 0.09), highlighting a POST improvement of RTG for the 20-mLB (∆=-7.9%; ES=0.85) and 20-mCoDB (∆=-5.9%; ES=0.33). In contrast, no improvements emerged for the SSG. The present findings indicate that the training approach of the RTG is more able to improve prepubescent soccer players’ sprint performances than that of the SSG, with the emphasis on ball possession executions, which are particularly game-related.