The aim of the present study was to examine the effect of cold water immersion (CWI) on sprint swimming performance in simulated competition conditions. Ten well-trained swimmers (5 males, 5 females; 19.0 +/- 3.9 years) performed two 100-m swimming sprints (S1 and S2) interspersed with a 30-min passive recovery period, during which athletes were randomly assigned to 5 min of CWI (14 degrees C) or an out-of-water control condition (CON 28 degrees C). During tests, sprint times, heart rate (HR), pre- and post-race parasympathetic activity via HR variability (natural logarithm of the square root of the mean of the sum of the squares of differences between adjacent normal R-R intervals; Ln rMSSD) and blood lactate accumulation ([La](ac)) and clearance ([La](cle)) were recorded. Rates of perceived recovery (RPR) and exertion (RPE) were evaluated before and after each sprint. CWI was associated with a 'likely' decrease in swimming performance [1.8% (90% CI 0.2, 3.5)], as well as 'likely' lower peak HR [-1.9% (-3.6, -0.2)]. CWI was also associated with a 'likely' smaller decrease in Ln rMSSD after the first sprint [-16.7% (-30.9, -4.1)]. RPR was 'likely' better [+27.2% (-3.7, 68.0)] following CWI. 'unclear' effects were observed for [La](ac) [+24.7% (-13.4, 79.5)], [La](cle) [-7.6% (-24.2, 12.7)] or RPE [+2.0% (-12.3, 18.5)]. Following CWI, changes in sprint times were 'largely' correlated with changes in peak HR (r = 0.80). Despite a subjective perception of improved recovery following CWI, this recovery intervention resulted in slower swimming times in well-trained athletes swimming in simulated competition conditions.