The rapid progress in mass storage technology has made it possible for designersto implement large data storage systems for a variety of applications. One of the efficient waysto build large storage systems is to use RAIDs as basic storage modules. In general, the datacan be recovered in RAIDs only when one error occurs. But in large RAIDs systems, the faultprobability will increase when the number of disks increases, and the use of disks with big storagecapacity will cause the recovering time to prolong, thus the probability of the second disk’s faultwill increase. Therefore, it is necessary to develop methods to recover data when two or more errorshave occurred. In this paper, a fault tolerant scheme is proposed based on extended Reed-Solomoncode, a recovery procedure is designed to correct up to two errors which is implemented by softwareand hardware together, and the scheme is verified by computer simulation. In this scheme, only tworedundant disks are used to recover up to two disks’ fault. The encoding and decoding methods,and the implementation based on software and hardware are described. The application of thescheme in software RAIDs that are built in cluster computers are also described. Compared withthe existing methods such as EVENODD and DH, the proposed scheme has distinct improvementin implementation and redundancy.