The oncogenic beta-catenin/T-cell factor (TCF) signal is a common trigger inducing expressions of various cancer-related genes and is activated in various types of human malignancy. The aim of this study was to create an effective double-stranded DNA decoy that would interfere with endogenous TCF hyperactivity in tumor cells. We first established the TCF-activated model using nontumor human embryonic kidney 293 (HEK293) cells by introducing a beta-catenin cDNA. Based on a consensus TCF-binding sequence in the cyclin D1 and c-myc promoters, several double-stranded oligodeoxynucleotides were designed and tested for their ability to inhibit TCF activity in the HEK293 model. Among them, the 18-mer oligodeoxynucleotide stably formed double-stranded DNA and efficiently inhibited TCF activity. FITC-labeled oligodeoxynucleotide was efficiently incorporated into the nucleus at 6 hours and remained within cells for up to 72 to 96 hours. When compared with scrambled oligodeoxynucleotide, we found that the 18-mer TCF decoy significantly inhibited TCF activity and promoter activities of the downstream target genes, such as cyclin D1, c-myc, and matrix metalloproteinase 7 in HCT116 colon cancer cells. Reverse transcription-PCR assays indicated that mRNA expression of these genes decreased with treatment of the TCF decoy. Proliferation assay showed that the TCF decoy significantly inhibited growth of HCT116 tumor cells, but not of nontumor HEK293 cells. Our data provide evidence that the TCF decoy reduced both TCF activity and transcriptional activation of downstream target genes. Thus, this TCF decoy is potentially an efficient and nontoxic molecular targeting therapy for controlling malignant properties of cancer cells.