For more than fifty years, diverse disciplines relating to occupational health hazards have witnessed the emergence of theoretical models explaining the impact of work environment on health. The existing summaries (e.g., Cox, Griffiths, & Rial-Gonzalas, 2000; Neboit & Vezina, 2002) often classify this models according to their interactional or transactional approaches. This theoretical distinction does not allow for further differentiations of approaches. Furthermore, attention is usually given to the most well-known models (e.g., Karasek & Theorell, 1990; Siegrist, 1996...), eclipsing others just as rich conceptually. These different elements explain the proposed review here, pivoting on the models which provide details on the relationship between work environment, stress and health at work. Of the review’s two objectives, one aims to present theoretical models in a historical perspective, to highlight the derivations of each research trend. The second objective concerns the proposition of a meta-theoretical framework, in order to characterize the models and compare them systematically. A documented research project has begun, using data bases to classify different disciplines around health at work. Seventeen models were selected based on explicit criteria and described in accordance with mediator and moderator variables used, classifed into categories (three for mediators: perception, evaluation, short-term reactions; and five for moderators: such as socio-demographics characteristics, social network at work, environment outside of work. In short, over half the models include perceptual processes as mediators, whereas evaluative processes intervened in only three of the cases. As for moderators, individual characteristics prevail and are more frequently supported than those associated with the work environment. The papers concludes with a discussion of the limits of this conceptual framework.