The aim of this study was to compare the effects of repeated sprints (RSA) training and regular soccer training on Yo-Yo IR-1 and RSA performance (6 × 40 m shuttle sprints). Thirteen semi-professional female soccer players and nine amateur male soccer players were randomised into a repeated sprint group (RSG; n = 12) or a regular soccer training group (STG; n = 10). The RSG soccer players executed 3-4 sets of 4-6 repeated sprints (30 m with 180° directional changes) weekly during the last eight weeks of the in-season. In parallel, the STG soccer players performed low- to moderate intensity soccer training in form of technical or tactical skills. The RSG showed 15% improvement in Yo-Yo IR-1 (p = 0.04; ES = 1.83) and their mean RSA times were reduced by 1.5% (p = 0.02; ES = 0.89). No significant changes were found for the STG (Yo-Yo IR-1, p = 0.13; RSA, p = 0.49). Comparing the groups, greater improvements were observed in Yo-Yo IR-1 for the RSG (p = 0.02; ES = 1.15), but not for the RSA (p = 0.23; ES = -0.33). Similar training volumes and intensities (% of HFmax) were observed between the groups (p = 0.22 and p = 0.79). In conclusion, a weekly RSA session integrated into a regular soccer regime improved in-season RSA and Yo-Yo IR-1 performance compared to regular soccer training.