BACKGROUND: "When will I be able to play again?" is the most frequent question asked by injured athletes. Due to the complex nature of sports injury, deciding when an injured athlete may safely return to training is a critical and difficult decision. OBJECTIVE: To study if the Safe Multidimensional Algorithm for Return to Training (SMART) scores, applied before the release to full return to training after injury differs between football players who suffer a subsequent re-injury and football players who do not. METHOD: Seventy one male professional football players were prospectively monitored for injuries during two seasons. The SMART tool was applied in injured players with an absence amp;gt; 10 days. The injured player had to carry out 17 multidimensional tests included in the algorithm in his final days of the planned rehabilitation. The results of the SMART were compared between players who sustained re-injuries and those who did not. RESULTS: Fifty-five injuries with absence amp;gt; 10 days were recorded and re-injuries occurred in 12 of these cases (22%). There was a lower re-injury rate in players who presented a better recovery in pain (p amp;lt; 0.001), agility (RR 21.0, 95% CI: 2.0 to 213.2), advanced agility (RR 26.7, 95% CI: 4.9 to 142.8), anxiety (RR 8.6, 95% CI: 2.0 to 36.2), depression (RR 10.3, 95% CI: 1.5 to 65.7), self-perception (p amp;lt; 0.001), advanced skills mode (RR 20.5, 95% CI: 3.3 to 125.9) and group skills mode (p amp;lt; 0.001). CONCLUSIONS: A multidimensional approach of Return to Training that includes objective measures may indicate potential deficiencies in the recovery of injured players.