Functional electric stimulation (FES) is a common intervention to correct foot drop for patients after stroke. Due to the disturbances from internal time-varying muscle characteristics under electrical stimulation and external environmental uncertainties, most of the existing FES system used pre-set stimulation parameters and cannot achieve good gait performances during FES-assisted walking. Therefore, an adaptive FES control system, which used the iterative learning control to adjust the stimulation intensity based on kinematic data and a linear model to modulate the stimulation timing based on walking speed during FES-assisted treadmill walking, was designed and tested on ten patients with foot drop after stroke. In order to examine its orthotic effects, the kinematic data of the patients using the proposed control strategy were collected and compared with the data of the same patients walking using other three FES control strategies, including (1) constant pre-set stimulation intensity and timing, (2) constant pre-set stimulation intensity with speed-adaptive stimulation timing and (3) walking without FES intervention. The error between the maximum ankle dorsiflexion angle during swing phase and the target angle using the proposed control strategy was the smallest among the four conditions. Moreover, there was no significant difference in the ankle plantar flexion angle at the toe-off event and the maximum knee flexion angle during swing phase between the proposed control strategy and walking without FES. In summary, the proposed control strategy can improve FES-assisted walking performances through adaptive modulation of stimulation timing and intensity when coping with variation, and may have good potential in clinic.