Cards game is one of the most popular entertainments in Asia such as: China, Vietnam, Cambodia, etc. Card games are usually played in special events such as Lunar New Year or daily to rest and have fun with relatives and friends. In addition, non-player character behavior study is also an interesting field of study, which aims to improve the experience of players with computer games. Therefore, the project’s goal is to find the best component so that player can take advantage of the most common card game, Tiến lên, which means “go forward” or “climbing up”, in Vietnam. The project uses four AIs to present four common players the game would have they are aggressive, rational, patient and rational & patient players. The aggressive robot makes a play as long as it can and only pass its turns when it cannot play. The patient robot intends to firstly play low-ranking cards & pass its turns to save high-ranking cards. The patient robot plays aggressively when there is a player with few cards left or the patient robot itself has few cards to play. The rational robot decides to play or to pass based on its ability of observing the table and its ability of predicting the odd to win the game. The patient & rational robot will be the crystallization of patient and rational robots. Under the client-server architecture, four robots will present four human players to play with each other. The performance of robots in the game of 2 players, 3 players and 4 players will be recorded and be analyzed to identify the best component in the game.