Glycated hemoglobin (HbA1c) has been increasingly accepted as the gold standard for diabetes monitoring. In this study, Raman spectroscopy was tentatively employed for human hemoglobin (Hb) biochemical analysis aimed at developing a simple blood test for diabetes monitoring. Raman spectroscopy measurements were performed on hemoglobin samples of patients (n = 39) with confirmed diabetes and healthy volunteers (n = 37). The tentative assignments of the measured Raman bands were performed to compare the difference between these two groups. Meanwhile, principal component analysis (PCA) combined with linear discriminant analysis (LDA) were employed to develop effective diagnostic algorithms for classification between normal controls and patients with diabetes. As a result, the spectral features of these two groups demonstrated two distinct clusters with a sensitivity and specificity of 92.3% and 73%, respectively. Then the effectiveness of the diagnostic algorithm based on PCA-LDA technique was confirmed by receiver operating characteristic (ROC) curve. The area under the ROC curve was 0.92, indicating a good diagnostic result. In summary, our preliminary results demonstrate that proposing Raman spectroscopy can provide a significant potential for the noninvasive detection of diabetes.