Secretive marsh birds (SMBs) are important indicator species of coastal wetlands but are difficult to detect and monitor. In coastal Louisiana, an important stronghold for these species, climate and hydrological models predict that freshwater and intermediate marshes will expand in the next 50 years, while brackish marshes will shrink. We used a multi-species Bayesian hierarchical occupancy model to estimate detection and occupancy probabilities for 11 SMB species in low and high saline marshes using data from automated recording units at 33 sites in southwestern Louisiana from February–June 2012. A quadratic effect of Julian date, but not minimum daily temperature nor precipitation affected detection of SMB species. King Rail (Rallus elegans), American Bittern (Botaurus lentiginosus), Common Gallinule (Gallinula galeata), and Pied-billed Grebe (Podilymbus podiceps) occupied mainly freshwater and intermediate marshes. Clapper Rail (Rallus crepitans), Seaside Sparrow (Ammospiza maritima), and Sora (Porzana carolina) predominantly occupied brackish and salt marshes. American Coot (Fulica americana), Purple Gallinule (Porphyrio martinica), Least Bittern (Ixobrychus exilis), and Marsh Wren (Cistothorus palustris) occupied both low and high saline marshes, showing flexibility that could maintain populations of these species as marsh salinities change in the future. If the current distribution of SMB species persists as marsh availability changes under future conditions, populations of the 4 species we found in low saline marshes may increase, whereas populations of at least 2 species found primarily in high saline marshes may decrease. Our modeling indicates that automatic recording units can produce comparable detection probabilities to other studies using traditional SMB sampling methods.