Abstract: In this paper, a network architecture based on multi-MEC (mobile edge computing) cooperative caching and streaming is proposed to address the problem that the delivery latency of mobile virtual reality (VR) video is too high under traditional delivery architecture. First of all, considering that the mobile VR video needs computing processing when watching and the mobile VR device may not always have the corresponding processing capacity, the MEC in the transmission architecture undertakes the needed computing task. Then, a mixed-integer linear programming (MILP) cache model with the optimization goal of minimizing the average delivery latency is established, and the optimal cache placement for mobile VR videos is obtained through the commercial solver Gurobi. Finally, numerical simulation experiments show that the proposed caching and streaming strategy can effectively reduce the average delivery latency of mobile VR videos when compared with existing caching algorithms.