The aim of the current study was to explore the association between urinary glycosaminoglycans and the onset of acute respiratory distress syndrome (ARDS) in sepsis patients. We conducted a prospective cohort study of 49 sepsis patients and collected clinical characteristics, urine, and blood samples within 24 h of admission to the Surgical Intensive Care Unit (SICU) of Peking University Third Hospital. Samples from patients who did not develop ARDS were collected from day 1 to day 7 after SICU admission, while samples from patients who developed ARDS were collected from day 1 to the day of onset, forming time-series data. A liquid chromatography-mass spectrometry (LC-MS) system was used to determine urinary dermatan sulfate (DS), chondroitin sulfate (CS), and heparan sulfate (HS) concentrations; enzyme-linked immunosorbent assay (ELISA) and cytometric bead array (CBA) were used to measure plasma inflammatory factors. The endpoint was the onset of ARDS. We used generalized estimating equations to identify associations between urinary glycosaminoglycans and ARDS onset in sepsis patients. Twenty-two of the 49 sepsis patients developed ARDS, and 132 plasma samples and 132 urine samples were obtained. Between the ARDS and non-ARDS groups, sex, pneumonia, Acute Physiology and Chronic Health Evaluation II (APACHE II), Lung Injury Prediction Score (LIPS), and PaO2/FiO2 showed statistically significant differences. Generalized estimating equations showed that DS and tumor necrosis factor-α (TNF-α) were independently associated with ARDS onset in sepsis patients, but their generalizability across ARDS subtypes warrants further validation.