The paper proposes an algorithm for fast mining global maximum frequent itemsets(FMMFI) in distributed
database, which sets center node. FMMFI algorithm makes computer nodes compute local maximum frequent itemsets
independently with discover maximum frequent itemsets algorithm(DMFIA) algorithm and local FP-tree. Then the center
node collects data with other computer nodes and combines data. Finally, global maximum frequent itemsets are gained.
FMMFI require far less candidate itemsets and communication traffic by using the strategy of top-down. Theoretical analysis
and experimental results show the effectiveness of the FMMFI algorithm.