In 2015, it is assumed that a great part of the population needs more and more assistance due to ageing. On the other hand, older people want to stay as longer as possible in their own home. The solution is then to implement automation functions with sometime biomedical monitoring. The considered system is composed of one server connected to electronic devices (sensors, actuators) through different protocols (KNX, 868 MHz, Bluetooth Low Energy …) and an interface (remote or not) enables the user to interact with. The design and implementation of such system requires addressing several issues: to facilitate the collaboration between these experts that have different responsibilities in the process (family members, doctors, maybe occupational therapists, nursing staff, and system installers), be easily definable, adaptable to final user's needs and upgradable to take into account the improvement of interface and/or functionalities provided by the industrials. To facilitate the cooperation between these experts and to better take into account the person needs, we propose a Collaborative Workflow to both generate the control code and the human computer interface (HCI) from parameters provided by experts. The implementation of this workflow is based on Model Driven Engineering. Based and the use of components. It enables to generate from a simple representation of the environment and a description of the need, both control programs and an interface prototype to control the smart home. The main process is split into 3 phases: The definition phase enables to acquire user's specific concerns, the environment and interaction models, the junction phase enables to express the technical part of the system and the generation phase provides the final control and HCI programs. The different phases rely on specific meta-models and model transformations refine the models. As a result, the approach invites many persons to interact according to their expertise. The solution provides application parts in an adaptive and automated way. The conceptual and the technical parts, separately specified, are linked and can be interoperable, opening the way for deployment in a varied e-health context.