Advancing age is associated with a decrease in step length. In line with previous studies showing that older adults often overestimate their motor abilities, we investigate whether older adults overestimate the length of their first step during gait initiation. The underlying effect could be a failure to update the internal model of motor action as a function of age-related motor decline. : Without taking a step, community-dwelling older women (n = 22, age range: 68-87 years) and younger women (n 19, age range: 19-33 years) estimated the length of their first step for both preferred step length and largest step length, which were performed without endangerment. Thereafter, the participants performed real gait initiation for both types of steps. The estimated step lengths were compared to the actual step lengths. : Older adults judged their first step as larger than it was (mean error: 30% for the preferred step and 9% for the largest step). A fine-grained analysis showed that this effect mainly concerned those for whom an increased risk of falling was suspected. These older adults were also among those who performed the shortest steps, and they presented with a slight decrease in cognitive functioning. Younger participants underestimated their preferred step length. Overall, the estimates were more accurate for the largest steps than for the preferred-length steps. : Step length estimation revealed powerful evidence for overestimation in older adults. Those who overestimated step length presented with more signs of motor decline. While this result sustains the idea of an insufficient actualization of the motor-action model, the explanation also refers to more global appraisal processes. Further research should explore the relevance of this task as a clinical laboratory tool for assessing gait capacity and the risk of falling.