There is a lack of information on the effect of recovery modes following team sport matches. PURPOSE This study aimed at investigating the most appropriate recovery mode following futsal matches. METHODS Ten futsal players (age: 23 ± 2 yrs; body mass 73 ± 7 kg; VO2max 52.2 ± 2.7 ml.kg-1·min-1; HRmax 196 ± 6 beat·min-1; Lamax 12.8 ± 2.6 mM) participated in this study. Data were collected twice a week for two weeks, when friendly matches were scheduled. Heart rates (HR), blood lactate (La), and rate of perceived exertion (RPE) were used to evaluate the intensity of the matches. At the end of the match, participants were randomly assigned one of the four 20-minute recovery protocols: 1) sitting rest (R); 2) low intensity dry-aerobic exercises (D); 3) low intensity shallow-water exercises (W); and 4) electrostimulation (E). Before, after the match, and after a 4-hour rest, photocells were used to measure athletes's counter movement jump (CMJ), bounce jumping (BJ), and 10m sprint performances. Salivary cortisol and athlete's leg muscle pain (MP) scales were collected before, at the end of the match, after a 4-hour rest, and the morning after the match. Urinary catecholamines were collected the morning of the match and the day after. The criteria for full recovery were afternoon test performances at least 97% of the morning pre-match ones. At the end of the experimental period, participants ranked the recovery modes by means of a 10-point likert scale. A Chi-square test was applied for heart rate counts and ANOVA was applied to test performances and hormonal values (p<0.05). RESULTS A preliminary analysis ascertained that morning pre-match test performances, cortisol, and MP values did not show any significant difference. Furthermore, no significant difference was shown between intensities (HR, La, and RPE) of the four matches. Cortisol values showed significant (F(3, 102) = 19.00 p <0.001) decreases during the day of the match and increases the morning after the match. Although catecholamine values were higher the morning after the match, no significant difference was found. After the game significant reductions for CMJ (F(2, 72) = 9.2 p <0.001) and 10m (F(2, 72) = 3.6 p <0.05) performances were found with no significant difference between recovery modes. However, D tended to show the best percentages of full recovery (range 60–100%). Recovery modes showed significant differences (F(3, 36) = 5.3 p <0.01) in the likert scale (E = 7.8 ± 1.4; W = 7.6 ± 2.1; D = 6.6 ± 1.8; and R = 5.2 ± 0.8). CONCLUSION Although recovery process seems not affected by active or passive modes, further studies are needed to fully elucidate the underlying mechanisms of recovery following training and match play.