/ Environmental quality assessment has to focus more on the quality of whole ecosystems, instead of focusing on the direct effects of a specific stressor, because of a more integrated environmental policy approach. Yet, how can the ecosystem quality be measured? Partly this is a normative question, a question of what is considered good and bad. At the same time, it is a scientific question, dealing with the problem of how the state of a system as complex as an ecosystem could be measured. Measuring all abiotic and biotic components, not to mention their many relationships, is not feasible. In this article we review several approaches dealing with this scientific question, Three approaches are distinguished; they differ in type of variable set and ecosystem model used. As a result of this, the information about the state of the ecosystem differs: ultimate breadth, comprising information about the whole ecosystem, is at the expense of detail, while ultimate detail is at the expense of breadth. We discuss whether the resultant quality assessments differ in character and are therefore suitable to answer different policy questions.How can the state of an ecosystem be described for environmental quality assessment? Just measuring all abiotic and biotic components, not to mention their many relationships, is not feasible. Several studies deal with this problem. They produce statements on environmental quality at an ecosystem level, i.e., ecosystem quality. These studies have been carried out for a variety of ecosystem types on different spatial scales and from different normative points of view. They have in common that the environment is regarded as a system of interrelated abiotic and biotic COmponents, rather than as a collection of individual components. For instance, the fauna species composition is regarded as a response to vegetation structure and abiotic habitat factors. For this reason, they are designated here as ecosystem approaches. In spite of this common character of the studies, different sets of variables and different types of ecosystem models are used to describe the state of the ecosystem. This re- view article investigates whether the variation in the ecosystem approaches in environmental quality assessment requires different kinds of ecological knowledge and data and whether the resultant quality assessments differ in character and are therefore suitable for different policy situations. The need for environmental quality assessment at an ecosystem level has been grown over the years. Initially, environmental quality assessment was oriented towards specific stressors and their direct environmental effects. Either load variables and concentrations of pollutants (e.g., Brown and others 1972) or biotic indicators for pollutants (e.g., Cairns 1974) were measured. The assessment focused on the question of whether a particular environmental problem was decreasing or increasing in severity. This approach was in line with the general strategy then followed in environmental policy making, in which ...