You don't know the location of the gene, you have no annotation information. The genome you have probably in form of contigs. That means you have to scan the whole collection of contigs and find candidate ORFs or conserved peptide domains. You seems to prefer spending little time on in silico analysis with "easy" ORF search and go straight to design and run PCR and all other stuffs. I am afraid that will take more time in total than do careful in silico analysis first. Even when you have done careful in silico analysis, you still need to choose several possibilities and design several primer sets to cover them. This will save your time from repeating the cycle several times. On the other hand, "easy" and "best" may not come together in this case. So, again it is rewarding to spend some time reading and trying established tools.
With that said, I think here is a quick ORF scanner tool which is easy to use:
http://star.mit.edu/orf/index.html
And here is an not so easy but good tool in my opinion:
Thanks for your effort Vang Le & Laid Douidah, But I tried with ORF finder method you told and I tried with many set of degenerative primers. Now i am trying with adapter based strategy. I hope I may get expected result.