Display omitted •Different pattern of the density distribution within pre-diabetic and diabetic liver.•The distribution of DAG composition dissimilar in pre-diabetic and diabetic liver.•The lipids species in the serum defined by MeV TOF-SIMS for the first time.•Serum analysis by MeV TOF-SIMS useful for in vivo investigation over time. Hepatic insulin resistance, induced by fat, occurs before peripheral resistance and leads to prediabetes and diabetes. If insulin resistance is detected earlier, lifestyle changes could prevent or delay disease development. Therefore, we analysed lipids in the liver and serum of prediabetic and diabetic mice by MeV TOF-SIMS with a focus on diacylglycerols (DAGs) as the best predictor of (liver) resistance. Glucose impairment was spontaneously developed or induced by HFD in NOD/LtJ mice, and prediabetic and diabetic mice were selected according to their glucose levels. MeV TOF-SIMS was applied to image the lipid distribution in the liver and to relatively quantify lipids related to insulin resistance in both the liver and serum. The same lipids were detected in the liver and serum but with different intensities between mice. The intensity of DAGs and fatty acids was higher in the diabetic than that in the prediabetic liver. Imaging of liver tissue showed a more compact density of prediabetic (non-fatty) than diabetic liver with DAG remodelling in diabetes. DAGs, which are greatly increased in diabetic serum, were successfully detected and quantified already in prediabetes. MeV TOF-SIMS applied to the serum presents an excellent tool for in vivo monitoring of disease development over time.