why to use zero ohm resistor, why not simply make a short circuit with a solder?
It can be used as a fuse for circuit protection under some circumstances...
. . or to bridge a track or to fit it where another version of the circuit needs a non-zero resistor.
Russell.
Because it can be used in production on a pick and place machine. Fitting or none fitting of a zero ohm link can allow different circuit configurations to be evaluated during development. It can also provide a tap in point if other things are required.
For example suppose you are not sure if some sub circuit needs an inductor or not in the decoupling circuit. You fit a zero ohm resistor and try it. If during testing you discover an inductor is needed then you simply fit one and change the bill of materials to include the inductor and exclude the zero ohm resistor.
Mind you, get the 1% zero ohm resistors, the 10% ones are rubbish. 
But the 10% offers more wiggle room for negative resistance, meaning more free energy.
...free energy.
Really?
Watcher:
Really?
No everything including and past my last sentence in my last post is a joke.
Grumpy_Mike:
No everything including and past my last sentence in my last post is a joke.
I do suspect "Watcher" has been around just long enough to figure that out. 
There are two principal reasons for using zero ohm resistors.
One is as an option - for example the common LCD modules from eBay and DX etc. tend to have two resistors in series with the LED, one being "101" or 100 Ohms, the other zero. One might ask why have the second one, but it may be easier to use a second resistor in some models and always place the "101". Perhaps more sensibly, some versions have places for four in a "bridge" configuration with only two zero Ohm jumpers ever fitted, which two determines which polarity is connected.
The other is where it is just convenient to run one or two tracks underneath, rather than using more "vias". In fact, this is frequently required on single-sided boards.
A lot of interesting answers.
In the manufacture of surface mount devices, the surfaces of
solder are protected from oxidizing by flux. This keeps from forming
a hard crust that keeps the solder from forming onto the leads, making
a good electrical connection.
It is desired to form this connection by surface tension.
When you have two pads next to each other, it is desired that they don't
normally short, with the help of both surface tension and flux.
So, if you want to have a method of intentionally shorting the pads that
normally would not connect, you need a part that will do the shorting.
That is why they have 0 ohm resistors.
If you can get a blob of solder to bridge the gap by piling on solder, it
will work as good as a 0 ohm resistor. It is not a method that would
work well in the manufacture of the boards.
Dwight
I often use them as jumpers over traces, and in setting up addresses ex. I2C