This paper introduced a new evolutionary algorithm — adaptive immune evolutionary algorithm (AIEA) for restoration optimization of all-optical network (AON), aim to design the virtual topology, globally optimize re-routing and throughput in the AON. By introducing adjusting parameters, the process of the algorithm can be adapted to the AON dynamically, so as to improve throughput and avoid premature or local convergence. Besides, the algorithm considers optimization of resource utilization in AON at the same time so as to decrease the blocking probability when failures happen in the networks. The results show AIEA has the best optimized performance in performance of optimization restoration in AON of MESH topology networks.