The food security problem in Africa is tied to the small farmer, whose subsistence farming relies heavily on rain-fed agriculture. A dry spell lasting two to three weeks can cause a significant yield reduction. A small-scale irrigation scheme from small-capacity ponds can alleviate this problem. This solution would require a water harvest mechanism at a farm level. In this study, we looked at the feasibility of implementing such a water harvest mechanism in drought prone parts of Africa. A water balance study was conducted at different watershed levels. Runoff (watershed yield) was estimated using the SCS curve number technique and satellite derived rainfall estimates (RFE). Watersheds were delineated from the Africa-wide HYDRO-1K digital elevation model (DEM) data set in a GIS environment. Annual runoff volumes that can potentially be stored in a pond during storm events were estimated as the product of the watershed area and runoff excess estimated from the SCS Curve Number method. Estimates were made for seepage and net evaporation losses. A series of water harvest index maps were developed based on a combination of factors that took into account the availability of runoff, evaporation losses, population density, and the required watershed size needed to fill a small storage reservoir that can be used to alleviate water stress during a crop growing season. This study presents Africa-wide water-harvest index maps that could be used for conducting feasibility studies at a regional scale in assessing the relative differences in runoff potential between regions for the possibility of using ponds as a water management tool. ?? 2004 American Society of Agricultural Engineers.