Researchers need to use complex models to understand socio-ecosystems (SES). However, observing SES models becomes a difficult task and no general framework exists to solve this problem. The aim of this paper is to propose a generic framework to specify the observation of SES models in order to easily generate the indicators that thematicians wish to monitor during simulation. To that end, we propose to reformulate the observation process as a transformation problem between data structures. This enables the concepts of Model Driven Engineering (MDE) to be used in the implementation of the domain specific languages (DSL). We designed dedicated languages to enable specify of what to observe, define which observation strategy to use, and decide how to generate the indicators.