With the petroleum exploration development toward high-density, high accuracy, high resolution seismic technology, seismic data processing and interpretation are facing data storage challenges. Storage virtualization is used to satisfy the demand of storage resource centralized management, sharing of seismic data between heterogeneous platforms, enhanceing the utilization of storage systems, I/O dynamic load balancing. This paper gives a design of storage virtualization by combination of storage device level and storage network level. Storage virtualization of logical level volume and file system level are applied in seismic data processing and interpretation storage system. The advantage is analyzed and verified.