Abstract: Background The fall risk for elderly individuals when stepping over obstacles of different heights with the dominant or non-dominant leg is unclear. Researching this issue can provide reference for the development of fall prevention measures for the elderly. Methods Sixteen elderly individuals randomly stepped over obstacles of 15%, 30%, and 45% leg height with both the dominant and non-dominant legs. A three-dimensional motion capture system and a three-dimensional force plate were used to record kinematic and kinetic data. Results When elderly individuals used the non-dominant leg to step over obstacles: (1) The dynamic stability in the anteroposterior direction of the trailing leg at 30% leg height was significantly lower than that at 15% leg height (P < 0.05, Cohen’s D = 0.628); (2) The vertical toe clearance of the trailing leg was significantly smaller than that when using the dominant leg (P = 0.041, Cohen's d = 0.516), while the vertical toe clearance of the leading leg (P < 0.001, Cohen's d = 0.685) and the anteroposterior toe clearance of the leading leg (P = 0.043, Cohen's d = 0.616) were significantly greater than those when using the dominant leg. Conclusion Compared to using the dominant leg, elderly individuals have a greater risk of falling in the anteroposterior direction when stepping over obstacles with the non-dominant leg; when stepping over obstacles of different heights, the risk of the trailing leg tripping over the obstacle is greater.