Many studies have shown that aerobic exercise improves cognitive function and maintains brain health. In particular, moderate-intensity exercise is effective for improving cognitive performance. However, there is no strong consensus on whether a single exercise session improves working memory (WM) function, as it does inhibitory function. It is possible that these discrepancies involve inter-individual differences in WM function. Therefore, we investigated whether acute mild and moderate aerobic exercise improve WM, and whether there exist inter-individual differences in improvements in WM. Thirty healthy subjects were recruited and participated in three experimental conditions (control, mild-intensity exercise, and moderate-intensity exercise). Subjects performed 10 min of exercise on a cycle ergometer with an individualized load. Their pedaling rate was maintained at 60 rpm. In the control condition, subjects rested on the cycle ergometer instead of performing exercise. The N-back task (2-back and 0-back task) was performed to assess WM function before, 5 min, and 15 min after the 10-min exercise session. In this study, to elucidate the effect of an acute bout of mild or moderate exercise on WM, the "2-back- 0-back" contrast, which is assumed to represent WM function, was calculated. The Two-Dimensional Mood Scale was adopted to measure changes in psychological mood states efficiently. The results revealed that working memory function was not improved by acute mild or moderate exercise. However, baseline working memory function was significantly associated with any change in working memory function following exercise, and this was independent of exercise intensity. Subjects with the lowest working memory function at baseline responded the most favorably. The results revealed that improvements in working memory function after a single session of aerobic exercise depend on baseline working memory function.