An interactive simulation system with a stereoscopic image and a method for operating the system are provided. The system includes a three-dimensional display and a control host. The stereoscopic image display is used to display a stereoscopic image. The system includes an interactive sensor that is used to sense gesture of a user or an action that the user manipulates a haptic device so as to produce sensing data. In the meantime, the system detects eye positions of the user. The changes of coordinates with respect to the gesture or the haptic device can be determined and referred to for forming a manipulation track. An interactive instruction can be determined. The stereoscopic image can be updated according to the interactive instruction and the eye positions of the user.