A serological survey of Visna/maedi virus (VMV) infection involving 274,048 sheep from 554 flocks was undertaken during 2002-2007 in Aragón, North-East Spain. One hundred and two of these flocks enrolled in a VMV control programme to reduce seroprevalence by selecting replacement lambs from seronegative dams and gradual culling of seropositive sheep. Twenty-five flocks were also visited to collect flock management and housing data. All study flocks had seropositive animals and 52.8% of animals tested were seropositive. Among flocks that joined the control programme 66 adopted the proposed measures and reduced seroprevalence significantly by between 26.1% and 76.9% whereas the remaining 36 flocks did not apply the measures and seroprevalence significantly increased. Seroprevalence increased with flock size and the number of days the sheep were housed, and decreased with increasing weaning age and shed open area, suggesting a reduced risk of VMV infection in sheep associated with better ventilation. At the end of the period, 24 flocks were certified as VMV-controlled with a seroprevalence <5%, and seven as VMV-free with 0% seroprevalence. These are the first officially recognised VMV-free flocks in Spain and represent a nucleus of VMV-free replacement animals for other flocks. Moreover, they are evidence of the possibility of eliminating VMV infection without resorting to whole-flock segregation or culling of seropositive sheep.