Abstract: AMS (Alpha Magnetic Spectrometer) is an international coorperative physics experiment with great influence and significance, and the AMS-02 tracker will be launched at the beginning of 2011. AMS-02 experiment will produce more than 1PB of massive data that have different types and their handling processes arc very complex This paper firstly introduced the functional requirements and the key components of the massive data processing environment of AMS-02 experiment,and then analyzed the fundamental data types and the handling processes of AMS-02 massive data,finally,paid more attention to AMS-02 SOC, the distributed architecture and the principal function of AMS-02 SOC were discussed, and a Grid-based implementation framework of AMS-02 SOC@SEU was also dicsribed.