The worldwide spread of severe acute respiratory syndrome (SARS) raised questions about the risk of importation of such infection, in particular by air travel. Entry screening was implemented in some countries although poor evidence on its effectiveness is reported. We developed a model to estimate the number of imported SARS cases between regions, using the 2003 SARS epidemic data to apply this model for two scenarios: from Beijing to Frankfurt and from Hong Kong to London. We back-calculated the data to estimate individuals' time of infection and built a model where every individual has a probability of being isolated, of traveling, and of being undetected at arrival. The findings, consistent with what was observed in 2003, suggest that entry screening does not affect the predicted number of imported cases. Inversely, importation depends on the transmission dynamic in the country of origin (including control measures in place) and on the intensity of air travel between regions.