Population monitoring is fundamental to the conservation of threatened species. This study aimed to develop an effective approach for long-term monitoring of the yellow-bellied glider (Petaurus australis) in north-east New South Wales. We conducted repeat surveys to account for imperfect detection and used counts in abundance modelling to produce indices of abundance. We used simulations to explore refinements to our study design. Surveys over three consecutive years produced 195 detections with >95% of detections by call. The probability of detection varied across years and survey occasions, ranging from 0.22 to 0.71. Abundance estimates were remarkably constant across years, ranging from 2.3 ± 0.5 to 2.4 ± 0.6 individuals per site. Occupancy estimates were also constant across years (0.90–0.91). Simulations were run to investigate the influence of the number of surveys (2 or 3) and the number of survey sites (20, 40 or 50) on the probability of occupancy. The design that reduced bias and provided an adequate improvement to precision was that of three visits to 40 survey sites. This design should be adequate to detect a decline in population abundance. Further studies of this kind are needed to better understand the population dynamics of this species.