Goal-oriented human saccades were recorded under double-step paradigm. The stimuli consisted of either visual or auditory-visual bi-sensory targets. Eye movement data were analyzed based on a 3rd-order linear horizontal saccadic eye movement model, where the inputs to the muscle were agonist and antagonist active-state tensions that were described by pulse-slide-step waveforms with a post inhibitory rebound burst (PIRB) based on a timeoptimal controller. Parameter estimations were calculated using the system identification technique for saccade parameters and neural inputs. Saccade amplitude transition function (ATF) and response latency indicated the saccade programming mechanism. The responses were affected by when the second peripheral target was presented. The neural input estimations supported the responsive neuron populations in the superior colliculus under different stimulus conditions, and indicated the double-step visual or auditory-visual stimulus may affect the synchrony of neuron firing. The study described here expanded our previous work and further supported the muscle model as well as the theory of the time-optimal saccade controller under physiological constraints.