In this paper we study a downlink communication system of ultra-dense 5G small cellular networks UDN with a large number of small base stations SBSs and user equipment UEs and consider the effects of power control strategy and queue state information QSI . The probability of successful data transmission between SBSs and its service users and the energy consumption of SBSs transmission are studied. Considering the interaction between SBSs we model the problem as a dynamic stochastic game DSG . By using the super dense characteristics of SBSs DSG is further approximated as mean file game MFG . Then based on Lax-Friedrichs scheme and Lagrange relaxation method a finite difference method is developed to solve the corresponding MFG. Before data transmission through the solution each small base station can independently optimize its control strategy to minimize the cost without any information exchange with other SBS. Simulation results show that the proposed control strategy can effectively improve the successful probability of transmission and reduce the corresponding cost.