The general consensus seems to be that GotM games unfairly advantage Domination/Conquest victories, and disadvantage peaceful victories.
What about basing the score on the following factors to balance out the score:
- Adjusted Population
- Adjusted Territory
- Average Culture
- Technological Advancement
- World Opinion
Adjusted Population
The population score would be based on the total population of your civilization, though only cities that have been in your possession (out of conquest revolt) for at least 11 turns, with that city's population "score" being reduced by 10% for each turn less than 10 that it's been in your possession down to receiving no score for a city that's only been in your possession 1 turn. So a newly founded / conquered city would take 11 turns before its population counts fully toward the score.
For example, if a player has 20 cities of population 5 each (100 base points) but two of the cities were founded four turns ago, the adjusted score would be 90 (18 * 5) for the cities that have been in possession for a number of turns, plus 10 (2 * 5) modified to 3 (10 * 30%) for only being in possession for 3 complete turns, giving a total effective population score of 93.
This would prevent the milking that so drastically inflates some scores.
Adjusted Territory
This would follow the same pattern as the population, where a square's value appreciates over 10 turns of possession to reach its full value. This is also to balance out any milking that may occur.
Average Culture
This is calculated to represent the total number of culture points per point of population, using the adjusted population score above (since newly founded/conquered cities have no culture, they shouldn't bring the score down).
Technological Advancement
You receive a score based on the total number of beakers your civ has produced, with bonuses for each spaceship part completed.
World Opinion
This is a factor that would make playing Switzerland worthwhile, in that you start with 0 world opinion, and get points for each green + you have and lose them for each red -that each remaining civ gives you. In a conquest game, you'd get 0, but in a diplo game, you could likely get enough of a boost from this to offset a lack of population and territory.
Conclusion
Obviously, it would need to be balanced to determine exactly what score you'd get for each population point or what the spaceship bonuses would be, but I think that it could serve to balance the CivIV scoring model enough to be allow players to be competitive regardless of differences in playing style, and focusing purely, but more comprehensively, on the civilization itself.
of the player vs. that of all players in the game. Use that instead of pop if it would score high than the pop score for culture victories. Since culture also gain you land maybe keep land for these VCs.
and 
of the player vs. that of all players in the game. Again replacing pop and land scores if greater for SS VCs. 
. I think there are two problems, both of which you hinted at in what you said.
