Numerous lessons have been learned so far in controlling H5N1 avian influenza in Asia. Early detection of incursions of virus prevented establishment of the disease in several countries, notably Japan, South Korea, and Malaysia. In countries where detection of early cases was delayed, infection is endemic and has been for three or more years. Control measures implemented in these countries need to reflect this finding. Vaccination will continue to be one of the key measures used in these endemically infected countries. Used alone, vaccination will not result in elimination of H5N1 viruses from a country, but, if used correctly, it will markedly reduce the prevalence of and susceptibility to infection. Vaccination has already played a valuable role in reducing the adverse effects of H5N1 viruses. Mass culling also reduces the level of infection in infected areas. However, the long-term benefits are limited in endemically infected countries owing to the high probability of reinfection on restocking unless other measures are used in parallel. Full epidemiological studies have not been conducted in many infected countries. Nevertheless, it is recognized that the number of clinical cases does not truly reflect the levels of infection. Domestic ducks and large live poultry markets have played a key role in the persistence of infection, because they can be infected silently. In tackling this disease, countries should adopt integrated control programs using the combination of measures best suited to the local environment. All surveillance data should be shared, both positive and negative, and should include information on cases of infection and disease. Socioeconomic and ecological implications of all control measures should be assessed before implementation, especially the impact on the rural poor.