The most important factors which contribute to the efficiency of game-theoretical algorithms are time and game complexity. In this study, we have offered an elegant method to deal with high complexity of game theoretic multi-objective clustering methods in large-sized data sets. Here, we have developed a method which selects a subset of strategies from strategies profile for each player. In this case, the size of payoff matrices reduces significantly which has a remarkable impact on time complexity. Therefore, practical problems with more data are tractable with less computational complexity. Although strategies set may grow with increasing the number of data points, the presented model of strategy selection reduces the strategy space, considerably, where clusters are subdivided into several sub-clusters in each local game. The remarkable results demonstrate the efficiency of the presented approach in reducing computational complexity of the problem of concern.