The tropical forests of Indochina harbor a suite of globally threatened tropical mammal species. These species are difficult to detect, and subsequently understudied. Noninvasive camera trapping was used to survey terrestrial mammals from a protected area in southeastern Lao PDR (Xe Sap National Protected Area). The presence-absence of four mammals (mainland serow Capricornis milneedwardsii, muntjac Muntiacus spp., macaque Macaca spp., and wild pig Sus scrofa) was modeled in an occupancy framework thereby accounting for detection probabilities. Our goals were to establish baseline occupancy data to assist with biological monitoring and to better understand the factors influencing the distribution of the target species. Naïve occupancy, or the proportion of sites at which the target species was detected, was 0.58 for muntjac, 0.55 for macaque, 0.38 for wild pig, and 0.30 for serow. True occupancy estimates (Ψ ± SE) from top-ranked models was 0.79 ± 0.21 for macaque, 0.74 ± 0.13 for muntjac, 0.51 ± 0.13 for wild pig, and 0.48 ± 0.18 for serow. The results underscore the importance of accounting for imperfect detection rates when studying rare or elusive species. I included two site covariates (forest type and distance to nearest village) in the occupancy models. Estimating occupancy as a function of site covariates improved model performance and provided insight into landscape-level factors that affect species occurrence. In the top-ranked models, serow occupancy was higher in hill evergreen forest (HEGF) than semi-evergreen forest (SEGF). Muntjac occupancy was higher in areas further from villages. Macaque occupancy was higher in areas closer to villages. Wild pig occupancy was higher in areas further from villages and in HEGF. I recommend using an occupancy framework to analyze occurrence data for difficult-to-study tropical mammal species. The results highlight the importance of Xe Sap NPA for large mammal conservation in the region. / text