Recent research on the construct of emotion suggests the integration of a motivational dimension into the traditional two-dimension (subjective valence and physiological arousal) model. The motivational intensity of an emotional state should be taken into account while investigating the emotion-cognition relationship. This study examined how positive emotional states varying in motivational intensity influenced set shifting, after controlling the potential confounding impacts of physiological arousal. In Experiment 1, 155 volunteers performed a set-shifting task after being randomly assigned to five states: high- vs. low-motivating positive affect (interest vs. serenity), high- vs. low-motivating negative affect (disgust vs. anxiety), and neutral state. Eighty-five volunteers participated in Experiment 2, which further examined the effects of higher vs. lower degree of interest. Both experiments measured and compared participants' physiological arousal (blood pressure and pulse rate) under the normal and experimental conditions as the covariate. Results showed no difference in switching performance between the neutral and serenity groups. As compared with the neutral state, the high-motivating positive affect significantly increased set-switching reaction time costs, but reduced error rate costs; the higher the motivational intensity, the greater the time-costs impairment. This indicates a role of the high-motivating positive affect in regulating the balance between the flexible and stable cognitive control. Motivational intensity also modulated the effects of negative emotional states, i.e., disgust caused a larger increase in time costs than anxiety. Further exploration into neurobiological mechanisms that may mediate the emotional effects on set shifting is warranted.