As potential endogenous biomarkers, reactive carbonyl species (RCS) have gained abundant attention for monitoring oxidative and carbonyl stress. However, there is no accurate method to evaluate multiple RCS in biological samples. In this study, a 2,4-dinitrophenylhydrazine (DNPH) derivatization-based LC-MS method was developed and validated to quantitate eight RCS: malondialdehyde (MDA), acrolein (ACR), 4-hydroxy-2-nonenal (4-HNE), 4-oxo-2-nonenal (4-ONE), methylglyoxal (MGO), glyoxal (GO), 3-deoxyglucosone (3-DG), and 2-keto-d-glucose (2-Keto). Subsequently, the method was applied to assess the RCS in low fat (LF), high fat (HF), and HF plus rosemary extract (RE) diet-fed mouse samples. The quantitative results on RCS levels indicated that the HF diet significantly increased the total RCS levels in mouse urine, plasma, and kidney with an average rate of 280.69%, 153.87%, and 61.30%, respectively. The RE administration significantly inhibited the elevated RCS levels induced by the HF diet, especially for MDA, 4-ONE, 4-HNE, and 2-Keto in mouse plasma, and ACR and 2-Keto in mouse kidney. This is the first study to simultaneously measure eight RCS in biological samples and demonstrate that RE was able to eliminate the accumulation of the HF diet-induced RCS.