<p>Drowning diagnosis and postmortem submersion interval (PMSI) estimation are still major challenges in forensic practice. Our recent studies provided evidence that microbiota successions in multiple organs, including intestine, liver, and brain, were valuable indicators for PMSI estimation. Meanwhile, microbiota in the lung from corpses submerged for 3 days presented obvious difference between drowning and postmortem submersion. However, gaps exist in our understanding of how long this difference lasts and how the decomposer microbial community in the lung changes with progression of decomposition. Here, we characterized the postmortem microbiota in the lung of mice submerged for 0 to 14 days, which were drowned or sacrificed by CO₂ asphyxia. Our study revealed that most samples collected before 3 days postmortem were not qualified enough for sequencing. The microbiota in the lung was largely influenced by the microbes colonized in the aquatic environment. Differences in microbiota between drowning and postmortem submersion faded over decomposition and could be used for drowning diagnosis within 10 days postmortem. Meanwhile, 22 taxa with good discriminative ability were identified to establish the classification model for discriminating drowning and postmortem submersion (AUC = 0.92, accuracy = 81.25%) by random forest algorithm. Twenty other candidates exhibiting obviously temporal changes were selected for PMSI estimation, which yield satisfactory performance (mean absolute errors ± the standard error = 0.976 ± 0.189 d). Altogether, we provide further evidence that microbiota signature of the lung is a promising bioindicator for the forensic death investigations of decomposed bodies recovered from water.</p>