People usually develop different kinds of compensated gait in response to local function deficits, such as muscle weakness, spasticity in specific muscle groups, or joint stiffness, in order to overcome the falling risk factors. Compensated walking has been analysed empirically in the impaired gait analysis area. However, the compensation could be identified spatially and temporally. The stability and perturbation resistance of compensated walking have not been analysed quantitatively. In this research, a biomimetic human walking simulator was employed to model one individual paraplegic subject with plantarflexor spasticity. The pes equinus was expressed by biasing the outputs of plantarflexor neurons corresponding to the spastic muscles. Then, the compensatory mechanism was explored by adjusting the outputs of the other muscles. It was shown that this approach can be used for quantitative analysis of the spastic gait and compensated walking. Thus, this research can improve the understanding of the behaviour of compensated walking, bringing insights not only for building useful walking assist systems with high safety but also for designing effective rehabilitation interventions.