To synthetically and dynamically make strategic choices in social networks, a novel adaptive approach to deal with two-dimension influence maximization problem (TIMP) is proposed with game-based diffusion model, which can achieve trade-off between diffusion time and the number of active nodes. At first, TIMP model is synthetically formulated, and diffusion time and the number of active nodes are defined mathematically. In particular, budget efficiency is presented to describe TIMP in order that an appropriate trade-off between diffusion time and the number of active nodes can be reached. Then, an adaptive heuristic (A-Heuristic) scheme is proposed to dynamically determine the initial seed nodes (ie, the most influential nodes). Finally, experiments are performed, and their results verify the superior performance of the proposed scheme in terms of budget efficiency, running time, and the number of active nodes.