Government smallholder irrigation schemes (SIS) were developed in former homeland areas of South Africa during the apartheid era. These schemes are now facing serious problems, owing to deteriorating infrastructures, low yields, limited access to services, weak and unclear institutions regarding water and land, and lack of support. Although representing a small percentage of irrigated land at country level, their location in poor rural areas represents a potential for poverty alleviation and food security. The central and provincial governments aim to both revitalise SIS and curtail the financial burden of their maintenance and operation costs. South Africa has cautiously embarked on the transfer of a number of schemes from government agencies to water users associations in the Northern and Eastern Cape Provinces. It is difficult for decision-makers and operators to evaluate the potential for long-term sustainability, then to organise rehabilitation and transfer accordingly, owing to a context of low participation, weak local institutions, and lack of information regarding farmers' strategies, land tenure arrangements, cropping systems, household socio-economics, and so on. A research team from CIRAD (Centre de Cooperation Internationale en Recherche Agronomique pour le Developpement) and the University of Pretoria proposes an action-research approach of SIS, in three steps: (1) data collection on the socio-economic and technical circumstances at household and scheme level; (2) the development of a model that calculates both the costs incurred by scheme management, and the possible contributions by farmers to cover these costs; and (3) running the model on a scenario-testing basis, evaluating the impact of certain measures or decisions, or certain farmers' strategies. The following principles form the background of the approach: (1) establishing and sustaining multi-disciplinarity and partnership; (2) considering local and specific circumstances; (3) developing and