As a new form of Internet service, User Generated Content (UGC) community has gained popularity these years, and also attracts academic attention. The posting behavior of users is an important index to measure whether the users are active or not. Identifying the potential active users in UGC Communities helps the companies make proper management strategies, which renders the prediction of users’ posting behavior important. This study used Poisson regression to examine the effect of users’ past behavior and status on users’ later behavior of posting. The data for analysis were extracted from ten thousand users on Qiushibaike throughout a window of three months, which included the number of posts and the number of likes, comments, and times they were featured on the home page by administrators. The results indicate that users’ past comments have a negative effect on users’ later posting behavior; users’ past posts have a positive effect on users’ later posting behavior; users’ past likes and hot posts they got have a positive effect on users’ later posting behavior.