In freestyle swimming, the amplitude of velocity (V) variation of the centre of mass (CM) during certain phases of the stroke cycle (SC) has been linked to swimming performance. The main purpose of this study was to determine accurately, in three directions, the intracycle variations of the CM V in male freestyle swimmers. Eleven male swimmers of up to international level performed a maximum 200m freestyle swim. Results showed that average CM V decreased as the race progressed. The magnitude of decrease was not associated with swimming V. Noteworthy fluctuations were found in all three directions, with the magnitude not correlated to swimming V. Absolute values for intracycle V fluctuations in all directions generally decreased throughout the test. No differences were found however for the corresponding percentage values (fluctuations calculated as percentage of CM average horizontal V). An exception in both cases was the horizontal V fluctuation, which increased in SC3. This increase seemed to be mostly associated with drop in intracycle minimum – rather than maximum- instantaneous V. Although swimmers achieved high average V mostly by reaching high peak V, the results suggested that for maintaining technique efficiency by minimising intracycle fluctuations, sustaining a decrease in minimum V was equally important to maintaining high peak V. Swimmers were found to roll their shoulders and hips in phase, but shoulder roll was considerably larger than hip roll. Swimmers increased hip roll but maintained the amount of shoulder roll during the test, with faster swimmers tending to roll less than slower swimmers. Neither elbow nor knee angular ROM was found to be associated with V, suggesting that other parameters such as arm orientation/V and joint angular V should be examined. The data revealed lateral asymmetries in all parameters measured, emphasising the importance of bilateral measurement of kinematics. Swimmers were found to produce higher CM V during the right arm underwater phase of the stroke and to roll the shoulders more to the left. No consistency between-swimmers patterns were observed for the other kinematic parameters. It was also shown that swimmers achieved maximum V at the same part of the SC that shoulder/hip roll peaks were observed. Similarly, minimum V coincided with neutral shoulder/hip roll and maximum elbow flexion. These results suggested that future research could focus on kinematic and kinetic measurements that would allow identification of propulsive efficiency for the aforementioned positions of interest.