Alzheimer's disease (AD) is regarded as a progressive neurodegenerative dementia, characterized by degeneration of distinct neuronal populations. A case-control study was carried out using high-resolution mass spectrometry to explore AD-associated urinary metabolic biomarkers from 30 AD patients and 30 cognitively normal (CN) individuals. In total, 49 metabolites were determined and validated as known compounds using LC/MS analysis. Using the two-sample t-test statistical analysis (P < 0.05), 19 metabolites were shown to be significantly different from AD to CN. A diagnostic model of the receiver operating characteristic curve was constructed with a combination of nine molecules out of 19 metabolites, it yielded a separation with an area under the curve value of 0.976 between the two groups. This study indicated that urinary metabolites showed a significant expression between AD and CN. AD-related metabolites enable to satisfy the diagnostic power of disease discrimination. In addition, as a noninvasive approach, urine collection is done easily in clinical diagnosis of AD.