The differential evolution algorithm is robust, easy to use, and requires few control parameters. However, as to the local optimizing ability, it is limited. Based on the principium analysis of the algorithm, the adaptive modification of the cross rate and the cross operator is proposed to improve the efficiency of the algorithm. To enhance the local optimizing, the local enhanced operator and the disturbed operator are proposed. Numerical study is carried out using five benchmark functions. Compared with the PSO algorithm, DE and MPDE, the ADMPDE is the most efficient algorithm of all, which verifies that the modification is effective.