As to the relationship between extraverted disposition and preference for social communication on Internet community, there have been two hypotheses: "Social Enhancement"("Rich Get Richer") and "Social Compensation" ("Poor Get Richer"). In this paper, we examined the role of using real-life portrait as facial picture in determining whether an extraverted user would be sociable on Internet community. 203 Chinese Social Networking Site (SNS) users were administrated with Big Five Inventory (BFI) to collect their scores on extraversion, and their web use behaviors were downloaded via Application Programming Interfaces (APIs). Results showed that, the factor of using real-life portrait as facial picture played a moderating role in the relationship between scores on extraversion and number of online friends (beta = 0.33, p < 0.05). For people using real-life portrait as facial picture, increased scores on extraversion were associated with increased number of online friends (beta = 0.33, p < 0.001); conversely, for people using none real-life portrait, there were no significant relationship between scores on extraversion and number of online friends (beta = -0.04, p = 0.85).