Is guard column often used in most of the LC methods these days? Is it necessary to use guard column for a shorter column (such as 10 mm) or with a gradient mobile phase?

If your analytical column is 10mm, most any guard column is going to be longer than that. I use guards on my 150x4.6 and 250x4.6 columns to prevent junk buildup at the head of the column. But then again, I inject whole bloods processed thru SPE. I think that sample type determines the need for a guard (or lack thereof).

My thinking about guard columns is that they can perform two functions.
1) physical filtration - to remove particulate matter
2) chemical filtration - removes chemical species that might affect the analytical column's properties
The first function is best performed by an in-line filter (IMHO)between the injector and column.
Therefore if your samples are clean, no need to have a guard column!
With dirty samples either perform sample preparation or be prepared to replace the guard column regularly.

Don't forget that a guard column is *just* an extension of the analytical column, so ensure that the chemistry is the same in both. Retention times will increase proportionally to the percentage increase in total column volume.

An additional benefit I feel guard columns provide by filtering out particles, co-extractives that may precipitate, or other bad things that get in there at 2000 psi, is they seem to shield the analytical column from compression when the pressure goes up. We've been able to get unbelievable life from LC columns that have had guard columns attached, and replaced when needed.

We do almost exclusively phenyl,C8 and C18 gradient RP work, so I use the physically strongest, most inert spherical particle C18 I can buy for guard columns, and don't bother matching chemistries. Most of the time, it won't matter because they are so short, and the retention orders similar for our compounds of interest. If there is a problem, I just remove the guard to see if things change.

At least put a particle frit in! I even use one with a Micra 1.5 um column--cheap insurance. I have run dust samples in methanol, without filtration (stirred, not shaken, let it settle, pipette off a mL), with a frit and guard column. Changed the $5 frit every 20-30 samples, or when the system got to 280 bar. Data was excellent.

A guard column becomes part of the analytical system, so while it may provide chemical filtration for the analytical column, you will never know unless you run the analytical column with a new, or no, guard column. Your analysis will look the same, except for the additional resolution the guard column adds.