Normally it would be by using 16S as marker and do a search on BLAST or GreenGenes. However I would not call it species, but OTUs, since in my opinion species definition in Bacteria is a bit iffy. What you are trying to do is associated with metagenomics. I am attaching a couple of links you might find useful.

http://greengenes.lbl.gov/cgi-bin/JD_Tutorial/nph-16S.cgi

https://www.biostars.org/p/58279/

Thanks Shaheed, but this is completely suitable when you are targeting specific species. In my case, i want to define every species in the sample.

Dear Dr. Mohamed, 

A quick way is to perform multiplex PCR with different promers specific for each type. Also as a control step, one can repeat the multiplex PCR using single primer in the mixed sample, then you will detect each species alone. 

Thanks  Dr Mohamed, but what i need is to identify all species that present in a sample (i have no guess what species may be present) so as Jose Sergio Hleap replied me above, it is a metagenomic analysis work!!