OBJECTIVES: The aim of the present study was to investigate the effects of cold water immersion during post-exercise recovery, with different durations and temperatures, on heart rate variability indices. DESIGN: Hundred participants performed a protocol of jumps and a Wingate test, and immediately afterwards were immersed in cold water, according to the characteristics of each group (CG: control; G1: 5\' at 9+/-1 degrees C; G2: 5\' at 14+/-1 degrees C; G3: 15\' at 9+/-1 degrees C; G4: 15\' at 14+/-1 degrees C). METHODS: Analyses were performed at baseline, during the CWI recuperative technique (TRec) and 20, 30, 40, 50 and 60min post-exercise. The average HRV indices of all RR-intervals in each analysis period (MeanRR), standard deviation of normal RR-intervals (SDNN), square root of the mean of the sum of the squares of differences between adjacent RR-intervals (RMSSD), spectral components of very low frequency (VLF), low frequency (LF) and high frequency (HF), scatter of points perpendicular to the line of identity of the Poincare Plot (SD1) and scatter points along the line of identity (SD2) were assessed. RESULTS: Mean RR, VLF and LF presented an anticipated return to baseline values at all the intervention groups, but the same was observed for SDNN and SD2 only in the immersion for 15min at 14 degrees C group (G4). In addition, G4 presented higher values when compared to CG. CONCLUSIONS: These findings demonstrate that if the purpose of the recovery process is restoration of cardiac autonomic modulation, the technique is recommended, specifically for 15min at 14 degrees C.