Abstract: Mining frequent itemsets with item constraints is a key step for mining association rules with item constraints.The meaning of mining frequent itemsets with item constraints is discussed first，the paper affirms that an itemset X satisfy an item constraints B by itself is not enough.It can be said X satisfy B only if all transactions that support the X satisfy the B.Based on this point，the algorithm Direct is modified to be Direct*.A negative item is looked as an independent item in Direct*.Item constraint is succinct，while the algorithms proposed do not utilize the succinctness adequately till now.A new algorithm MSEB is proposed and it utilizes the succinctness adequately.Experiments show the correctness of the two algorithms：Direct* and MSEB，which also show the efficiency of MSEB.