I have a file containing a list of interested gene (human) names, for example, PRKAR1A, PRKCB (some are aliases). How can I get the ID of these genes? So that they are uniquely identified? I am using http://www.ncbi.nlm.nih.gov/gene/?term=PRKAR1A
Using Biomart as mentioned by @poisonAlien probably is your best bet.
That said, you could also use BridgeDb. If I understand you correctly, essentially what you want to do is to map two different database IDs for human genes where the first is the HGNC name (used by NCBI as the gene name) and the other is the NCBI gene ID.
You could do that with a BridgeDb webservice call like: