You can use different bioinformatic softwares such as mirTarBase, miRWalk DIANA-Tarbase and others. Most of these websites give experimentally validated miRNAs target genes.
I appreciate you answer, but my concern wasn't about target for each miRs rather than the interaction between miRs-miRs, specially when I have a different two panels of miRs
One approach would be, you make a list of target genes for each miRNAs and then identify similar miRNAs based on commonly shared (overlap) target genes ref: Article Wang Y, Li X, Hu H. Transcriptional regulation of co-express...
. For this, you can perform (a) statistical test (eg. hypergeometric test, http://mirtrail.bioinf.uni-sb.de/tutorial.php) and (b) q-test (eg. similar to the one here http://www.pnas.org/content/100/16/9440 ) on the p-values obtained from (a).