The prevalence of neuropsychiatric diseases, including anxiety disorders, has increased in recent years. A better understanding of the mechanisms mediating symptoms in these disorders is essential for developing treatments. Although voluntary exercise can alleviate symptoms, its anxiolytic effect varies with the intensity of the activity. Therefore, to investigate the usefulness of voluntary exercise in alleviating the symptoms of neuropsychiatric disorders, assessing its effect based on intensity is required. Hatano rats, consisting of high- and low-avoidance animals (HAA and LAA, respectively), differ in their propensity to voluntary exercise. These animals are useful for examining the effects of voluntary running activity differing in intensity on anxiety-like behavior. We housed Hatano rats in cages containing locked or unlocked running wheels starting at 4 weeks of age, conducted elevated plus maze test at 8 weeks of age, followed by plasma corticosterone measurement and DNA microarray analysis on hippocampal tissue at 9 weeks of age. Our results show that only LAA (mild-intensity running animals), but not HAA (high-intensity running animals), had reduced anxiety-like behavior without plasma corticosterone change. In addition, LAA had increased immunity-related gene expression, but decreased proteolysis-related gene expression. Our findings suggest that mild-intensity voluntary running mediates the anxiolytic effect of exercise and is regulated through increasing the expression of immunity-related genes or decreasing the expression of proteolysis-related genes in the hippocampus.