Preclinical animal model studies of brain energy metabolism and neuroinflammation in Alzheimer's disease have produced conflicting results, hampering both the elucidation of the underlying disease mechanism and the development of effective Alzheimer's disease therapies. Here, we aimed to quantify the relationship between brain energy metabolism and neuroinflammation in the APP/PS1-21 transgenic mouse model of Alzheimer's disease using longitudinal in vivo 18F-FDG and 18F-DPA-714) PET imaging and ex vivo brain autoradiography. APP/PS1-21 (TG, n = 9) and wild type control mice (WT, n = 9) were studied longitudinally every third month from age 6 to 15 months with 18F-FDG and 18F-DPA-714 with a one-week interval between the scans. Additional TG (n = 52) and WT (n = 29) mice were used for ex vivo studies. In vivo, the 18F-FDG SUVs were lower and the 18F-DPA-714 binding ratios relative to the cerebellum were higher in the TG mouse cortex and hippocampus than in WT mice at age 12 to 15 months (p < 0.05). The ex vivo cerebellum binding ratios supported the results of the in vivo 18F-DPA-714 studies but not the 18F-FDG studies. This longitudinal PET study demonstrated decreased energy metabolism and increased inflammation in the brains of APP/PS1-21 mice compared to WT mice.