It's not as much about diode, as about optics. You need single mode laser, that's for sure. But rest of the problem is on optics side. Generally speaking, to improve divergence of a laser, you need to increase it's output aperture size - in other words the beam must be thicker when exiting laser.
0.3m diameter at 600m means 0.5 mRad divergence. That does not sound that unreal.
to get 0.5 mRad you need this formula:
divergence = (2 * wavelength) / (aperture_diameter * pi)
from that you will get
aperture_diameter = (2 * wavelength) / (divergence * pi)
using wavelength 650nm (red, which is cheap), and divergence 0.5 mRad you'll get required output aperture of 0.8276 mm.
That is quite usual aperture size for modules we use for building lasers, or for hobbyist lasers sold everywhere.
Now aperture size is not only thing limiting divergence, there are other, like spherical aberration of the optics system. But it should be main factor and it should give you rough idea.
But as guys already said .. it won't be safe. Safe is 5mW. 5mW laser spread over spot 300mm in diameter will be visible in total darkness .. but that's it. Never in daylight, much less in full sun. In full sun even that 250mW might not be well visible at this spot size.
You could also use green lasers, which will have better visibility (roughly twice), but will be more expansive.
Maybe you could try one of these, and see how it's going ..
http://www.o-like.com/index.php?mai...oducts_id=18&zenid=3n6rib33i9f4m406v0pl0jhot3
http://www.o-like.com/index.php?mai...ducts_id=340&zenid=3n6rib33i9f4m406v0pl0jhot3