Objective: Given the high risk of falls among older adults with sarcopenia and the frequent need to perform cognitive or motor tasks during daily walking, the mechanisms through which dual-task conditions affect gait and dynamic stability in this population remain unclear. This study aims to compare older adults with sarcopenia to healthy older adults to investigate the effects of cognitive and motor dual tasks on gait characteristics and dynamic stability. The findings will help clarify the mechanisms by which dual-task performance influences fall risk and provide a theoretical basis for developing targeted fall prevention and intervention strategies. Methods: Forty participants were recruited, including 20 with sarcopenia and 20 healthy controls. A twelve-camera three-dimensional VICON motion analysis system, synchronized with two AMTI force platforms, was employed to collect gait parameters and margin of stability (MoS) during single-task (ST), motor-motor dual-task (MMDT), and cognitive-motor dual-task (CMDT) conditions. Kinematic and kinetic data were processed using Visual 3D software to compute MoS and gait parameters. A two-way repeated-measures ANOVA was performed to examine the effects of group and task on gait parameters and dynamic stability during walking. Results: Compared to the healthy control group, the sarcopenia group exhibited a significant decline in dynamic stability during dual-task conditions, manifested by reduced anterior-posterior Margin of Stability (MoS) (Sarcopenia: -0.162 ± 0.023 m; Control: -0.086 ± 0.042 m) and decreased center of mass velocity (Sarcopenia: 1.045 ± 0.149 m/s; Control: 1.144 ± 0.152 m/s) (P < 0.05). In terms of gait parameters, the sarcopenia group showed significantly shorter step length (Sarcopenia: 0.60 ± 0.03 m; Control: 0.64 ± 0.04 m) and reduced gait speed (Sarcopenia: 0.81 ± 0.04 m/s; Control: 1.24 ± 0.11 m/s), prolonged gait cycle duration (Sarcopenia: 1.31 ± 0.13 s; Control: 1.15 ± 0.11 s), and increased step width (Sarcopenia: 0.14 ± 0.03 m; Control: 0.12 ± 0.04 m) (P < 0.01). Compared to single-task (ST) walking, older adults with sarcopenia demonstrated a decrease in anterior-posterior MoS (ST: -0.121 ± 0.041 m; MMDT: -0.162 ± 0.023 m; CMDT: -0.169 ± 0.027 m) and center of mass velocity (ST: 1.183 ± 0.177 m/s; MMDT: 1.045 ± 0.149 m/s; CMDT: 1.045 ± 0.158 m/s), along with a compensatory increase in mediolateral MoS (ST: -0.025 ± 0.034 m; MMDT: 0.010 ± 0.063 m; CMDT: -0.021 ± 0.053 m). Concurrently, step length shortened (ST: 0.63 ± 0.03 m; MMDT: 0.60 ± 0.03 m; CMDT: 0.60 ± 0.03 m), gait speed significantly declined (ST: 1.23 ± 0.05 m/s; MMDT: 0.90 ± 0.05 m/s; CMDT: 0.81 ± 0.04 m/s), and both gait cycle duration (ST: 1.15 ± 0.14 s; MMDT: 1.31 ± 0.13 s; CMDT: 1.36 ± 0.17 s) and double support time (ST: 0.23 ± 0.04 s; MMDT: 0.27 ± 0.05 s; CMDT: 0.27 ± 0.05 s) were prolonged (P < 0.05). Notably, the cognitive-motor dual task (CMDT) had a more pronounced impact on postural control in the sarcopenia group, leading to significantly greater reductions in gait speed, mediolateral MoS, and center of mass velocity (P < 0.05). Conclusion: Elderly individuals with sarcopenia exhibit decreased stride length and gait speed, prolonged gait cycle and double support phase during dual-task walking, along with a significant reduction in lateral center-of-mass velocity under CMDT. Clinical assessments should prioritize evaluating the characteristics of gait and postural control in older adults with sarcopenia under dual-task conditions, aiming to optimize targeted postural control training programs and thereby reduce the risk of falls in this population.