Context. Precise and accurate methods are essential to assess wildlife populations for sound management. We surveyed a managed population of fallow deer (Dama dama) in a Mediterranean environment in Italy, where this non-native ungulate has a negative impact on biodiversity.

Aims. We compare nocturnal distance-sampling (deer are detected by thermal imagery at night) population estimates with demographic projections of the same population.

Methods. We estimated natural survival in fawns (0.86), yearlings (0.83), adult males (0.70) and adult females (0.90) using capture–mark–recapture. By integrating survival estimates with population structure, reproductive traits and harvest data, we performed demographic projections. We performed nocturnal distance sampling on foot by using a thermal imagery once a year (in autumn) from 2001 to 2005. We walked 75–77 km (71 transects) per each survey.

Key results. We showed that our survey design met distance-sampling assumptions. Distance sampling and demographic projections yielded similar and precise (12.6% < CV <24.1%) population estimates, showing a decreasing (–164.64 deer year^–1) population trend from 2755 deer in 2001 to 1877 in 2005.

Conclusions. We showed that nocturnal distance sampling is useful to monitor wild deer populations in forests effectively and that it represents a cost-effective tool to develop sounded management policy for this non-native species. We also provided, for the first time, a comprehensive stochastic population model for fallow deer in a Mediterranean environment. Using these population estimates, managers could reduce fallow deer population size to a level compatible with the conservation of the endangered Italian roe deer and improve forest regeneration.

Implications. Nocturnal distance sampling can be used to assess ungulate population living in dense forested habitats effectively and efficiently.