Objective To investigate the effects of muscular fatigue on knee joint proprioception. Design Prospective study. Setting Exercise physiology laboratory. Participants Sixteen (eight men and eight women) healthy volunteers ages 19–27 years, with no history of neuromuscular disorders, vestibular disorders, or lower limb injuries (e.g., ligament/meniscus tear). Intervention Three separate fatigue protocols [ramp test (RT), continuous test (CT). and interval test (IT)] were performed. All tests consisted of lower limb cycling on a computer-driven cycle ergometer (Lode). The RT was used to calculate the maximal aerobic power (VO2max) and determine the work rates for the CT and IT. Work rate for the RT increased 20/25 W/min to maximal exhaustion. The CT consisted of cycling at 80% VO2max until maximal exhaustion. The IT consisted of cycling alternately at 120% VO2max and at 40% VO2max for 30 s each to the point of maximal exhaustion. Main outcome measure In the standing position, subjects were instructed to perform a two-legged squat to specific knee flexion angles. The absolute angular error (AAE) was measured for each test angle using an electrogoniometer (Penny & Giles, Blackwood, Gwent, U.K.) placed laterally across the dominant knee joint. AAE was defined as the absolute difference between test angle and subject perceived angle of knee flexion. Results A statistically significant increase in AAE after the RT (1.0 ± 0.66°, p < 0.01), CT (0.70 ± 0.66°, p < 0.03), and IT (1.24 ± 0.79°, p < 0.01) protocols was observed in the male subjects. Female subjects reported a statistically significant increase in AAE after the CT (0.73 ± 0.73°, p < 0.03) and IT (1.1 ± 0.89°, p < 0.01) protocols and a nonsignificant increase in AAE (0.19 ± 0.70°, p > 0.5) after the RT protocol. Conclusion These findings suggest that exercising to fatigue may produce a change in subjects' reproduction ability of knee joint angles. This may represent a decline in proprioceptive function after heavy exercise bouts. Whether this suggested proprioceptive decline is at the clinical significance level (e.g., significantly altering joint stability and motion) cannot be determined from the present findings.