The aims of this study were to investigate the periods in which sprints occurred during official matches and analyze these sprints considering the effect of the playing position and different contextual variables. Electronic performance and tracking systems were used for the analysis of all sprints performed by players. Matches were recorded by video and synchronized with performance tracking data. A total of 252 sprints were analyzed. The greatest frequency of sprints was observed in the period 1 (0’–15’), followed by period 2 (15’–30’) and period 6 (75’–90’), regardless of the playing position (χ 2 = 31.35; p = 0.051). Most sprints were non-linear (non-linear sprints: 97.6%; linear sprints: 2.4%) and without ball possession (without ball possession: 95.2%; with ball possession: 4.8%) for all playing positions, but the role of the sprint and the field area in which the sprint occurred were dependent on the position (p < 0.001). Specifically, players covered ~17.55 m per sprint, starting at ~10.34 km/h, reaching ~26.74 km/h, maximally accelerating at ~2.73 m/s 2, and decelerating at ~3.61 m/s 2. Overall, the playing position and contextual variables had no significant effect on physical performance variables analyzed during these sprints. Therefore, this study allows performance practitioners to have a better understanding of when and how soccer players sprint in match-play. In this regard, this study presents some training and testing strategies that may be considered to improve performance and decrease injury risk.