I also suggest Targetscan. Targetscan conserved sites, or alternatively, look those that have smaller context score value. If you have available  array data-set for your tissue or cell type, you can reduce the number of targets and look only those that are expressed. In addition, you can subject your list to pathway analysis For instance http://biit.cs.ut.ee/gprofiler/    Those targets that are in the same pathway make usually more sense. Good luck!

There are a variety of online databases that provide published and predicted targets for miRNA sequences. Perhaps a good place to start.

For example: 

annotations: http://www.mirbase.org/search.shtml

or 

experimentally validates targets: http://mirtarbase.mbc.nctu.edu.tw/index.php

etc.... Google miRNA targets. You'll find plenty of resources to get you started, assuming I understood your aim.

Try miRecords at  http://c1.accurascience.com/miRecords/index.php

or targetscan at  http://www.targetscan.org/

you can find miRNA predict targets for Homo Sapiens, Mus Musculus, Rattus Norvegicus, Drosophila Melanogaster e C. Elegans in http://www.microrna.org

"http://www.mirbase.org/search.shtml"  This website can find target gene based on miRNA.

I suggest you to look at this paper. it will be helpful.

good luck...

Yes, there are many databses to predict miRNA target sites. we have used the following and worked well for us.

psRNATarget: a plant small RNA target analysis server

For details, pl. refer our article

Prasath D, Karthika R, Habeeba NT, Suraby EJ, Rosana OB, et al. (2014) Comparison of the Transcriptomes of Ginger (Zingiber officinale Rosc.) and Mango Ginger (Curcuma amada Roxb.) in Response to the Bacterial Wilt Infection. PLoS ONE 9(6): e99731. doi:10.1371/journal.pone.0099731

I recommend TargetScan 

http://www.targetscan.org

I also suggest Targetscan. Targetscan conserved sites, or alternatively, look those that have smaller context score value. If you have available  array data-set for your tissue or cell type, you can reduce the number of targets and look only those that are expressed. In addition, you can subject your list to pathway analysis For instance http://biit.cs.ut.ee/gprofiler/    Those targets that are in the same pathway make usually more sense. Good luck!

Other than the servers mentioned by Prof. Rebane you can try : PicTar (pictar.mdc-berlin.de), MicroCosm version 5 (www.ebi.ac.uk/enright-srv/microcosm), miRanda (www.microrna.org), DianamicroT version 3.0 (diana.cslab.ece.ntua.gr/microT) and miRWalk database (www.umm.uni-heidelberg.de/apps/zmf/mirwalk). Hope the information was useful.

I would like to suggest that you should use at least 4-5 independent miRNA target prediction sites and use only those targets which are verified by at least 3 different platforms you have used.

You can go through this paper..

"A guide for miRNA Target Prediction and Analysis using web Based Application"

Francisco J Enuita

if possible, you can go through my publication "Genetic variations in microRNAs and the risk and survival of renal cell cancer" in PubMed. You can find the detailed process of searching miRNA target binding site 

Important notice:

There are online tools to find miRNAs of two types: 1) gives you manually curated miRNAs (collected from papers from experiments proven in the lab), or 2) miRNA predication by computer algorithms.

So I think it is important that you know what direction you want. Online tools do mention if their miRNA are manually curated or computer predictions.