Social anxiety disorder (SAD) is characterized by persistent fear and avoidance of social contexts due to fears of evaluation by others. Previous studies have found that high social anxiety individuals reported less positive emotion and more negative emotion compared to low social anxiety counterparts. Thus, mounting studies have paid more attention on the emotion regulation in social anxiety individuals. Emotion regulation refers to the process that a person uses certain strategy to modify the emotional experience and expression in order to meet particular contextual demand. With the further development of this domain, theorists have increasingly emphasized the key role of flexibility in emotion regulation. To be specific, emotion regulation flexibility is conceptualized as the ability to adjust the selection and implementation of emotion regulation strategies as the context changes, and could be divided into flexible selection and flexible implementation. So far, several studies have investigated the flexible selection and flexible implementation in healthy individuals, and found that: (1) healthy individuals tend to choose engagement emotion regulation strategies (e.g. reappraisal) in low-intensity emotions, and disengagement emotion regulation strategies (e.g. distraction) in high-intensity emotions; (2) healthy individuals can both effectively utilize engagement and disengagement emotion regulation strategies in low-intensity emotions, whereas the effectiveness of engagement emotion regulation strategies is worse than that of disengagement emotion regulation strategies in high-intensity emotions. However, although the lack of emotion regulation flexibility has been assumed as the crucial cause of social anxiety, the flexible selection and flexible implementation of social anxiety individuals still remain unexplored. Relatedly, why social anxiety individuals have difficulty in regulating their emotions? Because of their inflexible selection, inflexible implementation, or the combination of them? In order to address this concern, the present study examined flexible selection and flexible implementation of 2 emotion regulation strategies (reappraisal or avoidance) in response to 2 emotionally evocative stimuli (social or nonsocial pictures) in 2 groups (high and low social anxiety college students). Firstly, reappraisal, as a typical engagement emotion regulation strategy, involves forming neutral reinterpretations that compete with emotional appraisals at a late semantic meaning stage. In contrast, avoidance, as a representative disengagement emotion regulation strategy, refers to replace existing emotional information and prevent incoming emotional information from being further processed. Secondly, social pictures contain the content of human face, while nonsocial pictures do not. Finally, based on the scores of The Liebowitz Social Anxiety Scale-Self Report (LSAS-SR), the high social anxiety group (n = 37) with a minimum LSAS score of 40 and low social anxiety group (n = 29) with a maximum LSAS score of 37 were selected. The study found that: (1) when faced with social pictures, high social anxiety participants choose significantly less reappraisal and more avoidance compared to their low social anxiety counterparts. However, when faced with nonsocial pictures, there were no significant differences between high and low social anxiety participants whether in reappraisal or avoidance. (2) When faced with social emotion pictures, there were no significant differences between high and low social anxiety participants whether in reappraisal or avoidance. Similarly, when faced with nonsocial emotion pictures, there were no significant differences between high and low social anxiety participants whether in reappraisal or avoidance. These results suggested that, social anxiety individuals have difficulty in regulating their emotions when faced with social emotion stimuli instead of non-social emotion stimuli. More importantly, inflexible selection, rather than inflexible implementation, might be a key feature of social anxiety individuals.