Objective: The danger of widespread and intense epidemics of yellow fever (YF) in Africa has become very serious, requiring urgent immunization response. Because it is not possible to vaccinate 100% of the adult population, the challenge is to prioritize immunization of the population at highest risk. An original risk assessment has been performed at the initiative of the World Health Organization, using modeling to enable countries to define populations currently at highest risk, which will be vaccinated in priority. Methods: Five exposure risk factors have been selected and collected at the district level in three African countries: Burkina Faso, Togo, Mali. The five indicators are: ecological risk zone, confirmed YF cases since 1960, suspected cases since 1960, number of years in which YF cases notified since 1960, district close to another district that has notified cases since 1960. A multi-criteria analysis based on multiple component analysis (MCA) has constructed a composite exposure indicator (CEi) from the five selected exposure risk factors. In reducing by mathematical projections the number of dimensions, MCA modeling synthesize complex data tables. Results: For each of the three target countries, three analyses have been done for rural districts, urban districts and rural + urban districts. Four risk clusters have been determined from the lowest risk to the highest risks, allowing the construction of detailed YF risk maps in Burkina-Faso, Togo and Mali. These “YF risk assessment maps” present in four colors the four risk clusters at each district level. Conclusion: This approach seem to be an original, robust and reproducible technique for risk assessment purpose, which can be applied to a number of diseases and technology assessment when the number of indicators (risk indicators, clinical indicators, biologic indicators, etc) make data interpretation, comparisons and decision making difficult.