The purpose of this study was to determine the reliability and validity of regulating exercise intensity by ratings of perceived exertion in step dance sessions. Ten male college-aged students voluntarily participated in 2 step dance sessions for 45 minutes at 70-80% of their heart rate (HR) reserves with a 1-week interval between sessions. The step dance sessions included the same choreography with 10 minutes of warm-up, 25 minutes of the main part, 5 minutes of calisthenics for legs and abdomen, and 5 minutes of cool-down. In each session, subjects' ratings of perceived exertion (RPEs) were determined by Borg's 6-20 scale together with HR and lactic acid (LA) levels with 10-minute intervals. Values for RPE, HR, and LA increased nonlinearly in both sessions, and their trends were explained by polynomial equations to the second degree. The RPE values increased throughout each session, whereas HR and LA showed a decrease in the last time interval, which indicated that RPE did not maintain exercise intensity at proper range. Reliability coefficients for RPE scores in the first and last session ranged from 0.602 to 0.684. These findings suggest that RPE was a reliable but not a valid method for regulating exercise intensity in step dance sessions.