The biggest issue with flitting about at high fractions of c is acceleration. For tactical combat, I highly doubt that missiles will be used at long enough ranges to make those sort of velocities. Yes, I know they can pull much higher accelerations then humans can, but how much higher? Even at 1000 m/s, it will take a missile 83 hours to reach c, during which it will cover about 41 light-hours. That's an extreme case, but there is a limit to how much acceleration an object can take, particularly when supported not by a more-or-less uniform field, but by an engine. For strategic combat, this is less of an issue, as you can launch from whatever range you choose.
There are two ways to limit the use of relativistic missiles. The first is to apply the relativistic rocket equation (
http://en.wikipedia.org/wiki/Relativistic_rocket) to objects near c. A good point to start applying that might be at exhaust velocities around .4c, with mass ratios above 1.5. This prevents the object in question from ever achieving lightspeed, or even cheaply getting to a good fraction thereof. For example, take an object with a mass ratio of e and an exhaust velocity of c. Conventional rocket science suggests that it will have a delta-V of c, but in actuality, it will only reach about .76c.
The second is to apply an accuracy penalty to missiles at high velocity relative to the target. This is to simulate that the missile simply doesn't have time to respond when closing at high fractions of the speed of light.