Main objective of this study was to propose an advanced method to evaluate one's physical activity by means of a portable device which is capable of long-term non-restrictive measurement. The device includes accelerometers and a barometer specially designed to catch features of vertical movements such as stair climbing. The embedded algorithm classified activity type and quantifies its exercise intensity in realtime. Usefulness of the proposed method was demonstrated in simple experiments. Healthy young and elderly subjects participated in the experiments. Two kinds of daily outdoor situations were considered in the experiment; mountain climbing and city walking. Experimental results showed that the method correctly classified all type of activities, and accomplished feasible energy consumption estimation in both situations. It was also suggested that the consideration of vertical movements made an improvement in energy consumption estimation, and then provided more precise and reliable estimation of physical activity.