The rising age of the European population brings increased costs in healthcare mainly related to the management of chronic diseases. Regular physical activity has been shown to help in the prevention and control of disease risk. Mobile phones have provided promising and emergent ways of promoting healthy lifestyles, allowing real-time monitoring and coaching to be delivered at any time and any place. The presented work adds new features to the Activity Coach, an ambulatory feedback system that aims to encourage physical activity. The Integral of the Modulus of Body Acceleration (IMA) is the unit used as an estimate for energy expenditure. Although previous research demonstrated the potential of this system, results also showed that adherence drops after a few weeks. The primary goal of this research was to design, implement, and test adaptive goal-setting and personalized feedback strategies in order to encourage physical activity. Regarding the self-adaptive goal-setting feature, the goal for each day is defined automatically based on the physical activity performed at that day of the week since the beginning of the intervention. Hence, the intention is to help the user to increase or maintain his level of physical activity taking his daily routine as a reference. The second element added to the system regards motivational feedback messages, a key factor in interventions aiming at behavior change. Based on the levels of self-efficacy, stage-of-change, and daily activity, the user is assigned to one of the six pre-defined feedback strategies in the system. The content of the motivational cues depends on the selected feedback strategy. The evaluation of the system focused on providing better understandable and more accurate feedback to the user. To do so, we evaluated the challenge and attainability of the goals provided to the user with (1) data acquired during previous studies, and (2) newly gathered data from a single-subject study. As part of the evaluation, we translated IMA counts into ‘steps’, a commonly understandable measure for physical activity, comparing the data acquired from the Activity Coach sensor with a Fitbit, a commercially available pedometer. Although further tests with more subjects and different activities should be performed, we suggest that the default values set to the system are in agreement with the Goal-Setting Theory providing challenging and attainable goals. The results from this research will be used in future experiments using the Activity Coach and can be adapted to other ambulatory feedback systems regarding promotion of physical activity.