The increasing concern for the environmental impact of human activities stimulates the development of new methods for the analysis and design of industrial processes. It is generally admitted that the considerations of environmental and other life cycle factors as early as possible in the design can contribute to a reduction of overall costs. This is illustrated by the so- called fish diagram (fig 1) which shows the evolution of the freedom of design and the corresponding evolution of the overall costs during the design process. Increased freedom of design allows the consideration of process integration alternatives like, for example, those based on energy integration. Knowledge of life cycle parameters from cradle to grave (fig 2), in particular for the choice of materials considered, generates a coherent environmental picture for a holistic evaluation of alternatives. An earlier consideration of these factors favors a cost effective evaluation of trade-offs between internal process improvements and post-treatments of effluents. For processes with significant time dependence of the major input and output parameters additional important questions for the designer are where and when to invest, in particular when staged or deferred cost investments are worth considering. Move towards more sustainable processes is favored by the improvement of: • design methodologies (holistic life cycle design and other process integration methods such as those who will be exposed in this paper) and • technologies or combination of technologies (integration of technologies) to achieve a given production. Process technologies are not fixed but, like living bodies, they tend to adapt to the decision making environment, which at present is mainly based on economic factors and regulatory issues. Ideally assessments at all levels (company, national, international) should be made in a coherent framework where all technology options can freely compete particularly when a major departure from present day economics can be envisaged, at least for sensitivity analyses.