Well, it might kind of depend on how they've implemented the GUI, which is a pretty important part of what OSX is, to make a quotable understatement.
Generally, in some 'Nix like environment, your work flow for this task might look like this:
figure out which program is being called when you clickly-clicky the "terminal" launcher/icon/widget.
figure out whether that program can be called with an additional parameter that would tell the terminal to run the mysql monitor.
create a launcher/icon/widget that calls the terminal program with the appropriate switches.
For example, in my *BSD GUIs, I often create a launcher that calls something like:
Eterm -b white -c white -f black -T MyServer--icon-name "MyServer (SSH)" -e ssh admin@somehost &
Of course, that would probably only help you if you are running Eterm, or another terminal which supports the "-e" (execute) option.
Some GUI's also support launchers that execute a program from within a terminal. So, you'd create a launcher/icon/widget that calls /usr/local/mysql/bin/mysql and tell it to run in a terminal (GNOME is this way, for example).
(edit: Just noticed that goldbug answered while I was in between tasks. That would work, too, but if you ever just want the terminal, you'll have to quit the mysql monitor first.... )
