Previous research has shown that activating a stereotype can influence subsequent behavior in a stereotype-consistent way. However, the mechanism of stereotype activation to behavior is not clear. Some researches agreed with ideomotor theory. They provided that priming effects result from a direct link between perception and behavior. Some other researches argued that the process between prime to behavior is more complex. Recently, a rapidly burgeoning body of research has developed that suggests the self can play an important role in affecting both the magnitude and direction of prime-to-behavior effects. In this paper, we propose that the activation of a stereotype may elicit higher performance via higher self-efficacy beliefs. In Experiment 1, At first, we let the participants (not enter the formal experiments) wrote down what kind of profession need the high IQ ,and what kind of profession doesn’t need the high IQ, 50 participants were randomly selected and assigned to the two experiment conditions. Participants in the experimental group were asked to describe typical behaviors and characteristics of professors. Participants in the control group were asked to describe typical behaviors and characteristics of cleaners. After the priming manipulation, participants were asked to answer the general knowledge test. Additionally, for each item, they had to indicate how confident they felt with their answer on a scale ranging from 1 (No unconfident) to 5 (very confident). We conducted a 2 (stereotype activation: athlete vs. loafer) X 2 (cognitive style: independence vs. dependence) between-subject design in experiment2. Cognitive style as a new variable came into our experiment. The depend variable was the Grip strength test scores and participants specific self-efficacy that was related to athletic persistence and self-control. The results of Experiment 1 showed that participants who had been primed with professors (experiment group) scored higher on the knowledge test than participants primed with cleaners (control experiment). After the mediation analysis, we got that self-efficacy reliably mediated the relationship between stereotype activation and performance in a general knowledge test. In the experiment 2, the main effect of stereotype activation was significant and the interaction effect between prime and cognitive style was obvious. Through the simple effect analysis, we got that the subjects who were field dependence are more likely to suffer the effect of stereotype activation, and the field independence subjects are less susceptible to this kind of situation. In the same way, the self-efficacy of one’s persistence also mediated the relationship between stereotype activation and performance in Grip strength test. Through the two experiments, we discussed the internal mechanism between the stereotype activation and the changed behavior. When activated the positive stereotype traits to participants, the positive stereotype traits would be include into individual’s self-concept, altered self-representations, and eventually changed the behavior.