Background. The World Health Organization recommends a minimum of 150 minutes of weekly moderate physical activity to stay healthy (WHO, 2010). There are multiple coaching applications on the market that can help achieve this objective, but their theoretical groundings often remain unclear. Existing research prototypes in virtual coaching do not always propose personalised human-computer interactions based on personality and well-being theories. Aims. Our goal is to define a model which can be used 1) to improve motivation to physical activity through the consideration of personality and motivation theories, and 2) to implement such a model in a mobile walking application and to evaluate it. Method. We leverage two theories: the Self Determination Theory (Ryan and Deci, 2008) and the Regulatory Focus Theory (Higgins, 1997). Both theories are used to inspire the design of our PADME model (Personality-based Adaptive moDel for Motivation in E-health). PADME follows the METUX design framework (Peters et al., 2018) and is structured along two main dimensions: “step goal adaptation” (e.g. daily goals adapted to user’s performance) and “motivational interactions adaptation” (e.g. motivational messages adapted to user’s personality in terms of regulatory focus: if she is “promotion-focused” or “prevention-focused”). Results. Our model is used to describe interactions as expressed in the literature or reported during experiments about physical activity coaching conducted in our lab. Conclusion. The PADME model is useful for describing personalised coaching interventions based on SDT and RFT. Future directions include its implementation in a mobile application for supporting personalised interventions.