Recently proposed wireless proactive caching has opened up a promising avenue for reducing the redundant traffic load of the future wireless network. In addition to infrastructure updating, caching schemes for different wireless caching scenarios are also essential problems of concern. Along with the benefits brought by proactive caching, the cost of proactive caching should also be taken into account, since caching schemes can be influenced by different interest relevant parties in wireless networks. In this article, we discuss several typical caching scenarios and apply corresponding game theoretical models to illustrate how the selfishness of different parties may influence the overall wireless proactive caching. Efficient caching strategies can be designed by carefully considering the relations and interactions among these parties by using game theory. We also outline possible future research directions in this emerging area.