Background: We recently reported that tumor necrosis factor (TNF) signaling via the TNFR1 and TNFR2 receptors mediates the effects of long-term exercise on locomotion, cognition and anxiety, but not depressive-like behavior. We now investigated whether the TNF signaling via its receptors also mediates the effects of short-term exercise on cognition, anxiety and depressive-like behaviors. Methods: Thirteen-month-old C57BL/6 (WT), TNF−/−, TNFR1−/−, and TNFR2−/− mice were provided with 4 weeks of voluntary wheel running followed by behavioral testing using an established behavioral battery. Each genotype had a respective non-exercise control. Results: There was no interaction between genotype and exercise in any of the tests but the main effect of genotype, and not exercise, were found to be significant in the open field (OF), forced-swim test (FST) and Barnes maze (BM). In the OF, the control and exercise TNFR2−/− mice spent significantly less time in the inner zone than mice in the control and exercise WT and TNF−/− cohorts. In the FST, control and exercise WT mice showed significantly higher immobility time than their control and exercise TNF−/−, TNFR1−/− and TNFR2−/− cohorts. In the BM, the latency to escape over 4 days of training was significantly higher in all KO groups compared to WT, irrespective of exercise. Also, the latency to escape to the original location during the probe trial was higher for control and exercise WT compared to corresponding TNFR1−/− mice. In contrast, the latency to escape to the new location was lower for control and exercise WT compared to control and exercise TNFR1−/− and TNFR2−/− mice. The latency to escape to the new location in exercise groups was longer compared to control within all genotypes. Conclusion: While TNF signaling via the TNF receptors mediates cognition, anxiety and depressive-like behaviors independently, it does not mediate the effects of short-term exercise on these behaviors in middle-aged mice.