Obsolete pesticides have accumulated in almost every developing country or economy in transition over the past several decades. Public health and environmental authorities are eager to reduce health threats by removing and decontaminating stockpile sites, but there are many sites, cleanup can be costly, and public resources are scarce. Under these conditions, it seems sensible to develop a methodology for prioritizing sites and treating them sequentially, as budgetary resources permit. This paper presents a methodology that develops cleanup priority indices for Tunisia. The approach integrates information on populations at risk, their proximity to stockpiles, and the relative toxic hazards of the stockpiles. The robustness of this approach is tested by varying model parameters widely and testing for stability in the rank-ordering of results.