Parkinson’s disease (PD) is a neurodegenerative disorder by degeneration of dopamine neurons, affecting motor controls related to basal ganglia. Because severe movement disorders such as gait disturbances are often observed, evaluation from gait analysis is useful. From such a background, Coefficient of Variation (CV) and Detrended Fluctuation Analysis (DFA) comes to be used as one of the methods for analyzing the variability of the stride interval in recent years. However classification of the severity of PD by stride interval variabilityhas not been reached to practical use enough. In this paper, in order to clarify the difference in age and the severity of PD patients, variability of stride interval were analyzed by CV andDFA. As a first step, we performed analysis of stride interval in three minutes’ walk of 17 PD patients, 13 healthy elderly and 12 healthy young people. Particularly, we divided PD patients based on the Hoehn and Yahr (HY) scale into an HY2 group (n=9) and an HY3 group (n=8) in order to examine the relation to disease severity. Results indicate that CV seemed to distinguish PD patients from healthy people and that DFA fractal exponenttended to be related to the age and the disease severity. From these results, gait analysis using both CV and DFA is suggested to classify participants into healthy young, healthy elderly, HY2 and HY3 groups. For future direction, there are possibilities for seeing the progression from healthy people to PD patients.