The prosthetic knee is one of the main rehabilitation tools for an amputee to walk as normal. However, the proper functions of a prosthetic knee mechanism depend highly on the quality of input data in its synthesized process. Using skin markers of stereophotogrammetry, this research has performed various gait experiments including level and inclined walking, ascending and descending stairs, and passive knee movement. By the calibration of gait cycle peaks, experimental data are modified to prevent from the adverse effect of varied walking speeds and swing amplitudes on the variation and correlation analyses. Based on the study of gait cycle characteristics, it is hoped to provide better data for future synthesis of prosthetic knees to approach real human knee motion.