Compared with the conventional self-regulated EEG neurofeedback strategy, a closed-loop feedback strategy can provide a more precise and robust brain wave modulation. By detecting a participant’s instantaneous phase of EEG alpha oscillation and delivering phase-locked visual stimuli to the participant, a new closed-loop feedback paradigm, named phase-locked visual feedback modulation (PLVFM), has been developed to modulate his/her amplitude and frequency of alpha oscillation. However, the underlying neurodynamic mechanism of the PLVFM technique is still not clear, which limits the developments of precise and individualized applications of this new neural modulation technique. In this study, a neural dynamical model based on the limit cycle attractor has been proposed for alpha wave simulation to explore the neurodynamic mechanism of PLVFM. Results show that the simulated dynamic behaviors are consistent with the real results of online EEG modulation. The external stimuli at a specific phase change the instantaneous radius and phase of alpha oscillation. The repeated phase-locked stimuli stabilize the alpha oscillation in a new trajectory in the phase space and further induce the change of the amplitude and peak-frequency of alpha wave. The current study improves our understanding of the visual-modulated alpha wave, which is an important step towards precise modulation of EEG activity for the modulation of sensory and cognitive states.