The uncontrolled manifold (UCM) analysis has been demonstrated to be a powerful tool for understanding motor variability. The purpose of this study was to use the UCM analysis to investigate the effects of load carriage and fatigue on gait variability. Whole-body kinematic data during treadmill walking were collected from 12 healthy male participants when fatigue and load carriage were applied. The task-level variable for the UCM analysis was selected to be the whole-body COM. We chose to analyze the whole-body COM data at two important gait events: right heel contact and right toe off, and the UCM analysis was carried out in the sagittal and frontal planes, separately. The dependent measures were UCM variability measures and UCM ratio. Three-way ANOVA was performed to determine the main and interaction effects of back-carrying load, fatigue, and gait events on the dependent measures. The results showed that frontal UCM ratio significantly changed with the application of back-carrying load and fatigue, indicating that both factors had effects on motor performance in stabilizing the whole-body COM in the frontal plane. These findings can facilitate a better understanding of the nature of motor variability due to load carriage and fatigue.