You've just asked one of the defining questions of cancer chemotherapy, and one for which there is not a particularly satisfying answer currently.  Mechanisms for selectively targeting cancer cells follow a few paths:

1) Exploit the rapid uptake of materials by cancer.  This is the theory behind e.g. cisplatin - that cancer cells take in more materials, so will acquire a toxic concentration of the drug before healthy cells do.  This is also the theory behind e.g. some cancer imaging.

2) Exploit cancer-specific pathways.  This is essentially what your Asparaginase type approaches are doing - trying to deplete something that cancer cells need (asparagine) and which normal cells use relatively little of.  But of course asparaginase depletes systemic asparagine, so you're simply hoping that only the cancer cells die because of that.  Alternately, you can target proteins inside the cancer cells that they substantially over express, and hope that normal cells tolerate that inhibition.

3) Exploit cancer-specific membrane receptors.  This is becoming more popular, and is seen in e.g. cancer immunotherapies, or some nano-medicine, where you attempt to bind to a receptor that is highly expressed on the cancer cells of interest, and block its activity directly, or use that binding simply to direct your agents to the cells of interest.

4) Use prodrugs that only activate upon entrance to the cancer cells.  One of the more common mechanisms I've seen here is to use light-activated drugs, and essentially let the drug permeate the whole system, but then irradiate just the tumor to have only the tumor-bound drug (and that of the immediately surrounding tissue, probably) become activated and start killing the cells.

5) Use minimally soluble drugs.  If your drugs won't achieve wide systemic availability, you can inject them directly into the tumor, and hope that they won't get distributed too far from there.  Of course, using such minimally mobile drugs has problems all its own, too.

But yeah.  People have tried these mechanisms, and others, and we have yet to develop a real totally selective anti-cancer therapeutic.  Cancerous and non-cancerous tissues are, quite simply, just very similar.  It's very, very hard to selectively target one over the other, especially compared to targets like bacteria, viruses, etc. 

See, First of all decide on which cancer cells or protein you want to work.

get the protein from PDB and Dock the choice of your molecule according to literature study.

after doing docking study on specific study you will come to know that, if docking study shows good interaction with protein , you can do synthesis of that particular molecule.

you can do homology modleing also for finding good molecule.

Its always better do in-silico study before doing synthesis.

Many thanks to all of you guys and especial thanks to Dr. William P Katt for your rich elucidation!

I agree  with W.  P. Katt.  It would be most wonderful if  drugs were found that would specifically target  cancer cells and not  normal cells resulting in a cure.  At best for now it is difficult, made even more so by the fact that each individual and therefore each cancer reacts differently to a specific drug or drug regimen; a reaction which is sex related, age related, and a host of other genetic factors including how fast a drug is metabolized and cleared.