This paper proposes a method for robust hand detection for interactive touchless display using cascade classifier technique. A hardware system comprising a transparent display, two video cameras and a projector is assembled to generate on-display-surface object images for touchless display. A well-trained cascade of boosted classifiers is applied to detect the position of the hand in the object image. Using this method, accurate and robust hand detection for touchless display can be achieved. The detected hand trajectory information can be converted into mouse and keyboard inputs for interactive control and manipulation in virtual applications like virtual assembly and virtual design.