This study assessed the effect of swimming training on anxiety-like behaviors and corticosterone. Thirty adult male Wistar rats were randomly assigned to five study conditions: swimming training (ST); exposure to chronic mild stress (CS); exposure to chronic mild stress followed by swimming training (CS + ST); exposure to chronic mild stress followed by a recovery period (CS + recovery); control. The exercise training consisted of 60 min of swimming exercise per day, for five days a week, and four consecutive weeks. A chronic mild stress program (CMS) was applied for a period of four weeks. Anxiety-like behaviors were measured by open field test (OFT). The number of excrements and blood corticosterone were used as physiological parameters of anxiety. To assess corticosterone, blood samples were taken 48 h after the last session of experiments. Compared to other study conditions, the lowest anxiety-like behaviors and corticosterone concentrations were observed in the ST condition in unstressed rats. In stressed rats, as in the ST + CS group, swimming training probably reduced some anxiety behaviors, but the results showed increased corticosterone compared to control and CS + Recovery. Anxiety parameters and corticosterone concentrations were greatest in the CS condition. In the ST group, anxiety parameters were less than for the ST + CS group. In the CS + Recovery group, anxiety parameters were less than for the CS group. In summary, self-paced swimming training could attenuate some anxiety parameters in both stressed and non-stressed rats. The effect of swimming training in unstressed rats was more prominent than in stressed rats. In stressed rats, a period of recovery was more effective than swimming training in reducing corticosterone. Mechanisms of anxiety reduction other than cortisol should be investigated in future research.