OBJECTIVE: To determine the relation of hamstring and quadriceps muscle strength and imbalance to hamstring injury using a prospective observational cohort study METHOD: A total of 102 senior male Australian Rules footballers aged 22.2 (3.6) years were tested at the start of a football season. Maximum voluntary concentric and eccentric torque of the hamstring and quadriceps muscles of both legs was assessed using a Kin-Com isokinetic dynamometer at angular velocities of 60 and 180 degrees/second. Twelve (11.8%) players sustained clinically diagnosed hamstring strains which caused them to miss one or more matches over the ensuing season. RESULTS: There were no significant differences for any of the isokinetic variables comparing the injured and non-injured legs in players with unilateral hamstring strains (n=9). Neither the injured nor the non-injured leg of injured players differed from the mean of left and right legs in non-injured players for any isokinetic variable. The hamstring to opposite hamstring ratios also did not differ between injured and non-injured players. A hamstring to opposite hamstring ratio of less than 0.90 and a hamstring to quadriceps ratio of less than 0.60 were not associated with an increased risk of hamstring injury. A significantly greater percentage of players who sustained a hamstring strain reported a history of hamstring strain compared with non-injured players (p=0.02). However, this was not related to muscle weakness or imbalance. CONCLUSIONS: Isokinetic muscle strength testing was not able to directly discriminate Australian Rules football players at risk for a hamstring injury.