Legume-based technologies have potential to contribute to crop and livestock productivity, to improved human nutrition and to resource conservation in low input agriculture. Adoption of such technologies, however, lags far behind claimed potentials. There is a need to review present technical and methodological approaches to the design and transfer of these technologies by research and development organizations. Research needs to shift from detailed experimentation with a few promising technologies towards developing a basket of a large number of prototype technologies which match the diversity of target conditions in farmers' fields. Preparing the resulting information in a user-friendly database or expert system can guide the choice of appropriate technologies for target sites. Development needs to improve its technical capacity to characterize target sites and the underlying causes of constraints and its ability to identify appropriate prototype technologies for farmers. The evolving research-development scenario with a large number of non-governmental organizations operating at the local scale provides an opportunity to experiment with more appropriate legume-based technologies for a large number of diverse farming systems, if those involved in research and development respond to the new challenge.