Abstract: According to the feature of a large offline database based on mine fingerprint matching location algorithm，in order to more efficiently handle the offline database to obtain more accurate underground personnel positioning services，a cloud-based service model was proposed.With the use of the advantage of Hadoop technology architecture in the large data processing，the match positioning process was decentralized to the various nodes of a distributed cluster to parallel processing.By this method，the program of mine personnel positioning was designed and the reliability and efficiency of this program was demonstrated.

Key words: Cloud computing platform, Off-line data, Fingerprint-based localization, Hadoop, Distributed