The progress of Internet and computer science creates a huge opportunity in game industry, especially in online games. The differences between those expert online game players and those novice players in terms of performance are found. These differences influence how players interact with the game characters and their partners. Consequently, some novice players may quit from the game because they find the difficulty to join the expert player groups. In this study, we applied the card sorting technique to analyze the differences in mental models of different online game player groups with different levels of expertise. Results showed that significant differences between experts' and novices' mental models were found. By further analyzing these differences, we identified novice players’ shortcomings, and proposed some new game learning strategies to help novice players catch up with expert players, such as to gain knowledge of the game from a game forum on the web.