The invention discloses a method for designing high-sensitivity and high-specificity mismatch nucleic acid sequences. The method includes computing delta MFE (minimum free energy) values of mismatch sequences, target sequences and homologous interference sequences to determine candidate mismatch nucleic acid sequences. The method has the advantages that the shortcoming of dependence on experienceof mismatch nucleic acid sequence designs can be overcome by the aid of the method, mismatch nucleic acid sequences can be reduced to a great extent, particularly, workload for designing mismatch primers can be relieved to a great extent, a few candidate nucleic acid sequences can be quickly obtained from large quantities of mismatch nucleic acid sequences, and the method is favorable for determining the optimal nucleic acid sequences from the few candidate nucleic acid sequences; the high-sensitivity and high-specificity mismatch nucleic acid sequences still can keep high affinity with the target sequences with low concentration on the basis of mismatch primers designed by the aid of the method, and is poor in affinity with the homologous interference sequences, accordingly, the amplification sensitivity of the target sequences can be effectively improved, and the method can be effectively used for highly homologous nucleic acid sequences and can be particularly effectively used for miRNA [micro-RNA (ribonucleic acid)] amplification.