Implementation of file fault tolerance is the key for preventing data loss in cloud. However, cloud storage service providers may not offer the committed level, which results in that users may suffer data loss and economic loss. Existing inspection algorithms of testing of data fault tolerance in cloud have disadvantages such as spoofing attack of pre-fetch, low efficiency and poor practicality, which can not detect the foul behavior of cloud storage providers within a certain probability. To deal with the above problems, utilizing the difference of sequential access and random access, a remote testing algorithm of data fault tolerance in cloud named (Difference of Random and Sequential access Time) (DRST) is designed. The core idea is that the time of reading blocks of a file stored in order on a disk is much shorter than reading blocks of a file stored random on different disks. A strict theoretical proof and a in-depth performance analysis to the proposed scheme are carried out. The results show that the proposed scheme can accurately detect whether the cloud storage provider supplies clients with the committed level of fault tolerance.What’s more, the proposed scheme is much more efficient than the existing ones.