Abstract <p>Accurate monitoring of primate populations underpins effective conservation, allowing for the prioritisation of conservation actions and adaptive management of populations. The gibbons (Hylobatidae) are an especially threatened family that has proven challenging to survey accurately. The most common approach—the auditory survey—is subject to imperfect detection and measurement errors, which must be accounted for during analysis. We surveyed a population of Endangered northern yellow-cheeked gibbons <i>Nomascus annamensis</i> in a potential stronghold for the species: Virachey National Park, Cambodia. From 63 listening posts distributed across the park in 2018, we recorded 426 song bouts. Using acoustic spatial capture-recapture modelling, we accounted for the detection process and measurement errors during analysis. We also incorporated uncertainty in the identification of gibbon groups—typically neglected in gibbon surveys—using a bootstrapping approach. A threshold detection function received the strongest support, with detection probability dropping to 50% by ~800 m. Surveyors had to contend with multiple groups singing simultaneously; as a result, there was substantial error associated with bearing and distance measurements. Density was associated with canopy cover (quadratic response) and elevation (positive, linear response). Mean density across the park was 1.97 groups per km² for a total of 7,510 groups (95% confidence interval 4,859–10,099). This is the largest known <i>Nomascus</i> population and forms the core of a tri-border (Cambodia – Laos – Vietnam) population of &gt;8,000 groups. To protect this critical population, ongoing forest loss and degradation must be curbed and transboundary connectivity across the tri-border area maintained.</p> <p></p>