Stair descent is a challenging task in old age. This study firstly investigated lower extremity kinematics during stair descent in young (YOU) and healthy, community dwelling older adults (OLD). Secondly, the impact of an exercise training intervention on age-related differences in stair descent was assessed. At baseline, a motion analysis system was used to determine spatio-temporal gait variables and lower extremity kinematics as YOU (n=23, age=27+/-3 years) and OLD (n=34, age=73+/-4 years) descended a three step staircase. The older adults were then divided into training (TRA) and control (CON) groups. For 12 months, TRA performed resistance, aerobic, balance, and flexibility exercises under supervision in a class environment (twice per week) and unsupervised at home (once per week). CON carried on with normal daily activities. Following the intervention, baseline measurements were repeated in TRA and CON. At baseline, total descent, stride cycle, and single support times were longer in OLD than in YOU. In addition, sagittal plane knee motion was lower in OLD whilst frontal and transverse plane pelvis and hip motion were higher in OLD. Exercise training did not reduce the age-related differences observed. In conclusion healthy older adults perform stair descent at a slower speed and with greater motion outside the plane of progression than young adults. We found no evidence that these differences are reduced by generic exercise training, at least in non-frail older adults.