There are mainly two kinds of diseases in intermittent claudication. One is lumbar spinal canal stenosis (LSS) and the other is peripheral arterial disease (PAD). Differentiating LSS and PAD is a critical issue. Wrong differentiation might cause patient death. At small clinics and hospitals, simple and cheap differentiation system is required. Concerning this, this paper investigates walking motions of the patients. The subject with LED markers walks on the treadmill until she or he feels pain. We measure the walking motion by camera and track the LED markers. Treadmill enables to measure walking motion for a long time in a small space, and LED marker provides precious position of every joint in the walking. Then, we can get the information such as joint angle trajectory, hemi-foot step, stance and swing phases without any other sensors like foot switch or force plate. We compare walking motions of healthy person, LSS patients and PAD patients, find their features and 3 factors for disease differentiation; average bending angle of knee joint at the start of stance phase, average dorsiflexion angle of ankle joint, and average hemi-foot step length. From the results, we also show that 2 dimensional images of walking motion for several seconds are enough for deriving the factors. Then, we can construct the simple examination system for the disease differentiation.