terahertz simulation

You can model an antenna at a lower frequency and scale the results.
Simulations tend to fail at frequencies >50 GHz, so running into Terahertz may mean a trouble

I do not agree with this statement. The physics and math don't change above 50 GHz, and correct EM simulation is certainly possible. However, the user needs to be very careful is defining the correct physical properties (not always known at these frequencies). The user also needs a good understanding of ports/excitations, so that the simulation setup corresponds to the physical setup in measurement. All this is not trivial, but the problem is not the simulator itself.

ok, you said that CST is unable simulating in terahertz frequencies?

No, I did not say that. CST is fine for THz frequencies.

for obtain a terahertz radiation we need to illuminate the substrate with a laser beam,
how we can apply this laser beam in CST?
has CST an option for use of laser beam in simulation?

You asked about "simulating terahertz antenna". This does not require laser beam. You can use similar methods as known from GHz antenna simulations.

but to obtain THz radiation we should use laser beam to make a carrent that produce a radiation in THz band.
not to use waveguide port or discrete port?

The THz antenna simulations that I saw used wave ports or discrete ports. No laser, all electronic.

If you want to do optical simulations, have a look here:
https://www.cst.com/Applications/Cat...ptical-Devices

maybe it's true, because finally it's a current that will produce, thus maybe it's possible to use for example discrete port, but i saw that most of THz antenna simulated with another simulator, i can say i didn't saw any simulation THz antenna with CST or it's really few,(they use for example COMSOL, Silvaco, ...) because CST can't to model laser pulse.