Antisense compounds, various forms of nucleotides or their analogs, inhibit gene function both in vitro and in vivo. Although antisense compounds have been used extensively not only as a basic research tool but also as therapeutics for various diseases, one of the major problems is the difficulty of obtaining optimal sequences to inhibit specific gene functions. Although the terms "sequence-specificity" or "sequence-nonspecificity" are often used, there is no consensus as to how to define and quantitate such sequence specificity. In this review, we introduced hybridization simulation for designing optimal antisense sequences. Each candidate antisense oligonucleotide is assessed by calculating its hybridization energy against potential hybridization sites within the specified database (including GenBank) using a realistic nearest-neighbor thermodynamic model, taking into account mismatches. The specificity of each oligonucleotide is then quantitated by the number of potential cross-hybridizable genes and their degree of cross-hybridization. Furthermore, if antisense sequences exhibit a high potential for hairpin formation, they are not recommended even if they are highly specific. Therefore, to select antisense sequences, one should calculate all the potential factors for each candidate oligonucleotide such as length, location, specificity, hairpin potential, mRNA secondary structure, and dimer formation.