A challenge in physical activity research is how to accurately assess physical activity (PA) in free-living. The current methods of measuring PA and energy expenditure has limitations (John M. Jakicic et al., 2004). The SenseWear Pro Armband is a newly developed commercially available accelerometer based energy expenditure device. It is worn over the right triceps muscle and monitors various physiological and movement parameters. The purpose of this study is to grasp the energy expenditure distribution of SenseWear Pro Armband.measurments per minute. The distribution graphs – histograms – of energy expenditure per minute are two normal distributions, a bi-normal distribution, one disturbution for relatively low energy expenditure range and the other for relatively high energy expenditure range. The bi-normal distribution model was applied to the energy expenditure data as a distribution model. In this study, relative the energy expenditure level can be determined by comparing the distribution parameters: mean, standard deviation and proportion of each normal curve. Daily or weekly energy expenditure for individuals can be calculated with these distribution parameters. For this purpose, we need to know how many days' data are needed to get reliable energy expenditure information. The study's 5 subjects collected 21 days of data. There were no major differences between 21 days' and 7 days' data distributions, even the distribution of 3 days' data looked similar to the 21 day's data distribution. Three to seven days' data is enough to get a reliable energy expenditure distribution. For the first normal curve, the means of the energy expenditure per minute ranged from .943 to 1.201 calories and the standard deviation ranged from .046 to .113. For the second curve, the means ranged from 2.508 to 3.287 calories and the standard deviation from 0.739 to 1.127. The percentages of the first curve were 60 to 71%. The correlation of the means of two curves for five datasets was .864. This means that if the mean of the first curve increases, then that of the second curve increases. The correlation between the standard deviation of the first curve and the mean of the second curve was .779. When the variance of the first curve grows, the mean of the second curve become grows. In conclusion, the bi-normal distribution model was the proper distribution model for energy expenditure data measured with SenseWear Pro Armband. Keyword(s): assessment, health promotion, technology