I read many informations from many webs, all never give an answer with easy understand.
The reason you're getting lots of different answers is that there is no hard maximum current a particular cable can handle since it depends on the application.
You need to specify what physical limit you don't want to exceed for your "maximum" criterion. Do you want the wire not to rise above ambient temperature some amount? If so, what amount? If you can only tolerate a 10°C rise, then the maximum current will be less than if you can tolerate a 50°C rise, for example. Or maybe the limiting parameter for your application is voltage drop, which could make the maximum current quite different again.
AWG #16 copper wire has a resistance of 4.016 Ω per 1000 feet, or 4.016 mΩ/foot, or 13.18 mΩ/m. If you need to keep the voltage drop along 1 m of wire to 100 mV or less, then you can't push more than 7.59 A thru it. If you need to limit the power dissipation to 1 W/foot, then you can't push more than 15.8 A thru it. 1 W/foot would get noticably warm, but should not be dangerously so for most uses.
If you are using it for house wiring, then it becomes a legal matter and you simply look up the answer. One chart I looked at says the limit is 3.7 A for "power transmission". That was chosen to be very conservative so that some amount of degradation and screwups can happen and your house is unlikely to burn down due to overloaded wiring. Legal limits will vary by jurisdiction and which electrical code is being applied. You will have to consult the electrical code that applies to your area and circumstance to get the answer.
That same chart that shows 3.7 A as the maximum for power transmission also shows 22 A maximum for "chassis wiring". Presumably more temperature rise is acceptable inside a chassis.
So to get a straight answer, you have to tell us what you are really trying to do. There simply is no inherent maximum current until you get to where the copper melts. Even that depends on ambient temperature assumptions.
It is important to note that the current that can flow through a wire depends on a few factors:
allowable voltage drop. As mentioned in the answer by stevenvh, it depends on the length of wire you are using and the application in question
allowable temperature of the wire, and insulation
I've usually worked with the idea that a wire at 70 degrees is pretty damn hot. I believe a lot of the time the ratings for cable insulation sheath is rated to 100 degrees.
If you have 16 AWG wire with a silicon insulation you'll find its rating to be substantially higher than cheaper plastic insulation.
This table says 3.7 A. According to the same table AWG 16 has a resistance of 13.17 Ω per km, which means that at 3.7 A one meter of wire would cause a power loss of 180 mW, and a voltage drop of 48 mV.
PetPaulsen's nomograph doesn't show a value for AWG 16, but you can compare the values for AWG 22 with those of the table. The table gives much lower current values. That's because the wire's electrical insulation also acts as thermally insulation, whereas the PCB traces the nomograph is about can drain their heat to both the PCB's mass and the air, so that their temperature won't rise as much as for insulated wire.