Mobile Edge Computing (MEC) is regarded as a promising technology that migrates cloud computing platforms with computing and storage capabilities to the edge of the wireless access network, enabling rich applications and services in close proximity to the mobile users (MUs). Previous works on computation offloading only focus on limited resources, interference etc. and leave service caching for MUs with heterogeneous computation task and opportunistic networks out of consideration. In this paper, we perform a novel research on the computation offloading strategy taking account into service caching and Device-to-Device (D2D) communication, and introducing opportunistic networks in multi-access networks simultaneously. First, we denote the computation offloading model and formulate the offloading decision as a sequential game problem. In order to solve this problem, we then design a suboptimal algorithm based on game theory. Finally, the performance of the proposed algorithm is verified by comparing with other algorithms. Simulation results corroborate that the proposed computation offloading strategy can not only decrease the overall computation overhead efficiently, but also achieve a Nash equilibrium.