At present, aiming at the inseparable relationship among the physical activity, energy expenditure and human health (chronic diseases), the measurement of physical activity is quite important to human health. Now most of the study supports of using the three-axis accelerometer to measure the physical activity, but for walking or running, the existing testing instrument and method is accurately only for flat ground. The standard method of analysis of body accelerations cannot accurately predict the energy cost of uphill walking without an external measurement of the slope. Therefore, This paper proposed an regression method of integration of data and HRmax-to-HRrest ratio(HRindex) by collecting and analyzing exercising data. In this study, the subjects finished the treadmill walking experiments wearing tri-axial accelerometer and the heart rate belt, then the three axis accelerometer data was separated instead of using the sum of the three axis integration on regression analysis. It was concluded that the method had a good measuring precision for treadmill experiment with different inclines.