Transcranial magnetic stimulation (TMS) of human motor cortex typically evoked motor responses. TMS has failed to elicit eye movements in humans, whereas prolongations of saccadic latency have been reported with TMS. In previous studied we demonstrated that saccades can be abolished or saccadic trajectories can be changed through TMS in the 100 msec before saccade onset. This effect was especially marked when TMS was applied parietally. TMS never influenced a saccade in flight. Simulations of predictive experimental saccades that were impaired through TMS of the frontal or parietal cortex demonstrated especially that the dynamics of small saccades were markedly influenced, resulting in a significant decrease in acceleration and amplitude, or an almost complete inhibition. The impact of inhibition through TMS was critically dependent on timing: early TMS (-70 msec) had a much larger inhibitory effect than late TMS (-20 msec) on experimental saccades. Differential timing of TMS in influencing the cortical control signal is demonstrated through simulations of a reciprocally innervated eye movement model that paralleled empirically determined changes in eye movement dynamics after real TMS. There is a reasonable match between the model and the experimental data. We conclude that the inhibitory action of a presaccadic disturbance, such as a TMS pulse, on saccadic programming is inversely related to timing and amplitude of the predicted saccade.