MicroRNAs (miRNAs) are a group of small non-coding RNAs consisting of 18~24 nucleotides in length. Each miRNA is expected to bind a few hundreds of putative target mRNAs, thus inhibiting their translation into protein products mostly by degradation of targets. With its biogenesis extensively deciphered, miRNAs have been implicated in a variety of biological processes, including early development and cellular metabolism. In addition, dysregulation of miRNAs and subsequent alterations in the expression of its target molecules are thought to be linked to the pathophysiology of multiple human illnesses, including cancer. To establish the miRNA-target relationships important for developing a specific disease, it is critical to validate the putative targets of each miRNA suggested by computational methods in vivo. In this review, we will first discuss oncogenic and tumor-suppressive roles of miRNAs in human cancer and introduce computational methods to predict putative targets of miRNAs. Then, the value of Drosophila melanogaster as an alternative model system will be further discussed in studying human cancer and in validating the miRNA-target relationships in vivo. Finally, we will present a possibility of applying the mammals-to-Drosophila-to-mammals approach to study the roles of miRNAs and their targets in the pathophysiology of oral cancer, an intractable type of cancer with poor prognosis and survival rate.