Cycling power decreases substantially during a maximal cycling trial of just 30 s. It is not known whether movement patterns and joint powers produced at each joint decrease to a similar extent or if each joint exhibits an individual fatigue profile. Changes in movement patterns and/or joint powers associated with overall task fatigue could arise from several different mechanisms or from a complex interplay of these mechanisms. The purpose of this investigation was to determine the changes in movement and power at each joint during a fatiguing cycling trial. Thirteen trained cyclists performed a 30 s maximal cycling trial on an isokinetic cycle ergometer at 120 rpm. Pedal forces and limb kinematics were recorded. Joint powers were calculated using a sagittal plane inverse dynamics model and averaged for the initial, middle, and final three second intervals of the trial, and normalized to initial values. Relative ankle plantar flexion power was significantly less than all other joint actions at the middle interval (51±5% of initial power; p=0.013). Relative ankle plantar flexion power for the final interval (37±3%) was significantly less than the relative knee flexion and hip extension power (p=0.010). Relative knee extension power (41±5%) was significantly less than relative hip extension power (55±4%) during the final three second interval (p=0.045). Knee flexion power (47±5%) did not differ from relative hip extension power (p=0.06). These changes in power were accompanied by a decrease in time spent extending by each joint with fatigue (i.e., decreased duty cycle, p<0.03). While central mechanisms may have played a role across all joints, because the ankle fatigued more than the hip and knee joints, either peripheral muscle fatigue or changes in motor control strategies were identified as the potential mechanisms for joint-specific fatigue during a maximal 30 s cycling trial.