first of all You should be familiar with all of the failure of that product (mean if you want to design a mechanical component then you should know about its failure criteria)and how can you minimize the cost of your product. Actually designing process is not so simple as you think, it need experience and imagination.

hope this is answer of your question....

The design of any mechanical part or any thing in the universe starts with an idea, which we can get by preliminary brain storming.

The next step is to have a simple drawing how the object is going to look like.

The third step will be converting the 2D drawing into a 3D drawing using any rendering software like CATIA, Solidworks etc

The fourth step will be to analyse the generated 3D object under loads and forces that you think that the object designed is going to experience in action.

This step will be useful if there are any design changes in the object, keeping in mind that there are any braking stresses generated during analysis.

After all the modifications has been done, the final design is used to make a mould for production.

first of all u have to aware about the design criteria

factor of safety

stress theory(6 theories)

Have a look at BS7000, it shows various models of design. Also you may find the TRIZ technique useful. (theory of design)

If you have no design experience, then design something quickly build it and make mistakes, that is how you learn. Use the Demming cycle.

If you have some experience, then start with user requirements, from these generate a functional specification, then do a brain storm on how to meet the specification. Think of as many ways as you can of meeting the requirements. Then evaluate each one as honestly as you can and compare with the user requirements.

Choose one idea and expand it. Do some quick  analyses, may be stress, current consumption, cost etc. whatever is appropriate. Do not worry about accuracy at this stage. You are just trying to put the idea into the right "ball park" make guesses if you do not know exact numbers.

Then design the system and break it down into modules.

Detail the modules using CAD, or paper drawings, what ever you have the skills to do.

Then you can perform more accurate analyses or do tests on each module.

Integrate the modules into the final product and test against the user requirements.

Hope this helps.

Paul

This question applies to all of the engineering disciplines, I think. And I believe that perhaps the most important element in design is "necessity." ("Necessity is the mother of invention," and for me, that's certainly been the case.)

So you define the mission for this invention, and then the environment in which it must function, and that leads to design details and tradeoffs. Things like product lifespan, tolerance for detectable and undetectable errors, recovery from error or failures, cost tradeoffs, and so on.

Often you'll want to quantify things that are not easy to quantify, so you find ways. One example might be, instead of quantifying "availability," you might derive availability from mean time between failures and mean time to repair. Those two numbers might be easier to establish. Or tolerance for undetectable errors might have to be quantified as "mean time between undetectable errors has to be no less than the product's estimated lifespan" (depending on whether this is a safety-critical product).

My mentor years ago explained to me that the hard part of design is always that you have to design for failures. If you design only for success, you haven't done your job.

Design is an informed intelligent decision making process. 

If you want to learn about design, start from books. First of all, you needd to be good at theory. Start with any good book like Shigley's Mechanical Engineering Design or Machine Elements in Mechanical Design by Robert L. Mott. Also have a look at Machinery's Handbook. Once you are good at theory, move on to softwares like SolidWorks, CATIA, Creo Parametric etc.

Thanks to aal of you for answering my query. Specially Mr. Mudassar for giving the name of books. And you are right when i am good at theory then I'll be good designer. And again thanks to everyone...