I've actually done something similar before for an autopilot for a prototype I made.
What I did was to blend the two thrust values together. I'd add the maximum thrust in the direction to negate the relative velocity to the maximum thrust in the direction to accelerate/decelerate the ship. (I'd also blend in other thrust values for orienting the ship so the biggest thruster would point in the direction of the greatest need) Once all these thruster values were added up, I'd scale them all back so that no thruster would be running at more than 100% power.
I played with the blending values a bit, but it always seemed to work fine when the weights were all the same for different priorities with the potential exception of orientation which was better with a smaller weight. It worked well in cases when you were close to the target, but had a large difference in velocity as well as when you were far away with a small relative velocity. This was because the smaller factor was usually eliminated quickly, thus dropping it's effective blending value to 0 and giving priority to other factors. This was very cool to watch, especially when you assigned a group of other ships to fly in formation with yourself.