Recommender systems have emergedrecently in many aspects due to enormous amount of information or items existed in e-commerce or social network services. Users find it convenient if the service can recommend interesting items to them automatically. Past systems adopted users' preference or social connections for such recommendation. In this work, we will try to identify a game player's personality traits and use them to recommend games that may be interested by the player. The personality traits of a player is identified by applying a text mining process on textual contents related to the player. The same process is also applied on the comments of the games to identified the personality traits of games. Recommendation is then performed based on the similarity between the traits of the user and the game. We performed experiments on 63 players and 2050 games and obtained promising results which may demonstrate the plausibility of personality-based recommender systems.