Even users know the privacy risks of information disclosure, many still fail to adopt protection mechanisms or decrease information disclosure behavior on social network sites (SNSs). We explain this phenomenon by considering users' rational calculus, affection, as well as perception towards platform. The awareness towards privacy leaks gradually fades out with the time passing. Accordingly, we proposed that users' self-disclosure behaviors on SNSs are a decision-making process determined by individual rational, emotional, or situational factors. Users' choices of persuasion routes depend on psychological distance of privacy invasion events based on the perspective of construal level theory. Results from analyzing collected 241 usable records show three findings. First, users' decision of privacy disclosure is based on institutional trust and cognitive absorption, rather than perceived benefits. Second, cognitive absorption influenced by perceived playfulness is critical to privacy disclosure behavior. The influence of cognitive absorption on privacy disclosure varies from the time period of personal bad experience happened. Third, the influence of trust on privacy disclosure is getting more important when users have bad experience and when users have modified privacy settings.