Background <p>The mechanisms underlying gait function recovery in stroke remain uncertain. Biomechanical gait analysis has emerged as a promising approach to address this gap, offering essential information for developing tailored gait rehabilitation strategies in patients with stroke. However, few studies have investigated the gait biomechanics of dependent stroke ambulators, particularly for patients classified as level 3 in the Functional Ambulation Category (FAC), which refers to the ability to walk on a level surface under supervision.</p> Methods <p>This prospective observational study recruited twelve patients with chronic stroke with an onset duration of more than six months, along with six healthy adults. The patients with stroke were grouped into FAC level 3 (N.=6) or FAC level 4 (N.=6) based on their level of independence. All participants performed level walking along a 7-meter walkway while three-dimensional motion capture was used to assess gait biomechanics. Seven functional assessments, including the Berg Balance Scale and Trunk Impairment Scale, were also conducted in patients with stroke. The Kruskal–Wallis test and one-way analysis of variance were used to compare gait parameters among groups, followed by Mann–Whitney and independent t-tests for post-hoc analyses.</p> Results <p>Both stroke groups showed significant differences in biomechanical parameters compared to the healthy group (<i>p</i> &lt; 0.05). Compared to the FAC 4 group, the FAC 3 group exhibited significantly lower peak posterior ground reaction force on the affected side (<i>p</i> = 0.002); reduced hip range of motion (<i>p</i> = 0.047) and peak hip flexion moment (<i>p</i> = 0.044), and maximum knee flexion angle on the unaffected side (<i>p</i> = 0.026). Compared to the healthy group, the FAC 3 group demonstrated significantly reduced ankle range of motion on the affected side (<i>p</i> = 0.021), and lower maximum hip extension angle (<i>p</i> = 0.011), lower peak hip extension moment (<i>p</i> = 0.031) and peak ankle dorsiflexion moment lower maximum hip extension angle (<i>p</i> = 0.004) on the unaffected side, while no differences were observed between the FAC 4 and healthy groups (<i>p</i> &gt; 0.05).</p> Conclusions <p>Significant differences in biomechanical parameters, particularly those related to eccentric contraction in the proximal joints of the unaffected side, were observed between FAC 3 and FAC 4 groups. These disparities highlight the need for tailored gait rehabilitation strategies based on ambulation level.</p> Trial registration <p>Clinical Trial No. NCT05908994, Date of registration: 23/05/2023.</p>