This paper illustrates the need for a revival and renewal of landscape analysis in order to identify, evaluate and predict environmental change in environmental impact assessment (EIA) and development perspectives. An initial focus on the landscape, rather than on individual systems or processes, makes it possible to widen the scope of study, to assess change in different temporal and spatial perspectives and thereafter to converge on key issues of relevance for specific areas or development projects. This approach can be seen as a process rather than a method, which calls for intradisciplinary competence in data collection and evaluation as well as an interdisciplinary assessment capability. A combination of such scientific competence, local knowledge, and experiences of the local environment is used to widen the assessment perspectives and the prediction competence. The approach is illustrated by its application in two Tanzanian studies. The Southern Highland study emerged from two feasibility environmental impact assessments (EIAs) of proposed hydropower projects whilst the Babati study was initiated as a result of previous sectorial research on land management, which had to be analyzed in broader perspectives. In both cases, a need to define environmental baselines to assess land use and project related environmental change had been defined by different donor agencies. One conclusion from our study is, however, that there is no such thing as an environmental baseline, rather a baseline that has to be extended in different temporal and spatial perspectives to fully understand and predict environmental and related social change. This study can therefore be seen as a contribution to a new understanding of environmental change that is required for strategic environmental impact assessments and long-term natural resource-use planning.