Purpose Elevated body temperature might change glucose metabolism in human organs. The purpose of this study is to explore F-18-FDG distribution in febrile patients on the day of F-18-FDG PET/CT scanning and compare it with patients with a normal temperature. Procedures F-18-FDG PET/CT was performed on 69 febrile patients and 82 patients with a normal temperature. Patient sociodemographic data, blood glucose levels before PET/CT, body temperature on the day of the exam, and laboratory test results were collected. Maximal standard uptake values (SUVmax) in the brain, mediastinal blood pool, liver, spleen, and the bone marrow were compared. Results Compared with the controls, SUVmax of the febrile patients was significantly lower in the brain, mediastinal blood pool, and the liver (p < 0.01), and higher in the spleen and bone marrow (p < 0.01). In the febrile group, SUVmax was not significantly different between the FDG burden and non-FDG burden patients (p > 0.05). Body temperature was found negatively correlated with SUVmax in the brain (r = - 0.646), mediastinal blood pool (r = - 0.530), and the liver (r = - 0.384), and positively correlated with the SUVmax in the spleen (r = 0.592) and bone marrow (r = 0.651). Multivariate linear regression established body temperature on the day of PET/CT as an independent affecting factor (p < 0.01) for the SUVmax in the brain, mediastinal blood pool, liver, spleen, and bone marrow. The SUV in the brain, liver, and mediastinal blood pool remained different (p < 0.05) after corrected with the SUVmax in the blood pool or liver. Conclusions Fever influences F-18-FDG distribution in multiple human tissues and organs. Altered F-18-FDG distribution in vivo might affect results of disease lesion detection and tumor therapy response assessment. Correction with blood pool or liver SUV fails to cancel the effects of fever. The day of fever should be avoided for PET/CT scan, especially in assessing tumor therapy response.