Most human cases of highly pathogenic H5N1 avian influenza virus (HPAIV) infection are the result of direct contact with infected poultry. Therefore, infection of poultry should be prevented to avoid human exposure. One method to combat HPAIV outbreaks relies on depopulation. An alternative or supplementary method is the use of DIVA (discriminating infected from vaccinated animals) vaccines to prevent infection of animals on holdings surrounding an outbreak. Discrimination between infected and vaccinated animals is often based on the ‘heterologous neuraminidase’ strategy. This implies that a suitable vaccine can only be selected when the N-subtype of the outbreak strain is known. Thus, at least two vaccines with different N-subtypes must be available, allowing a switch of vaccine in the event that one of them matches the outbreak strain. However, such vaccines cannot be used preventively in situations in which the N-subtype of the outbreak strain is unknown. In order to circumvent these drawbacks we generated a recombinant influenza virus containing the HA gene of a contemporary H5N1 HPAIV strain in combination with the NA gene of a human type B influenza virus. An inactivated vaccine based on this virus protected chickens against clinical disease, and completely prevented virus shedding after H5N1 HPAIV challenge infection. Serological analyses confirmed that the vaccine complied with the DIVA principle. Since NA of type B does not occur in avian influenza strains, this vaccine is suitable as a DIVA vaccine against any H5 HPAIV, and may be used preventively without compromising the DIVA principle.