According to Travelling (LBB'77, p2) a "Typical Travel Time" is 9.6 days for 1 billion km under constant acceleration, so by that logic a ship would use less fuel than a smallcraft. I'm not quite buying that a 5000 Dt (or 100 Dt) ship uses less fuel than a 50 Dt smallcraft for the same acceleration...
Yes, there are 1008 turns in a week, but I wouldn't read too much into that, I guess that is more happy coincidence than cunning calculation.
A ship is not normally accelerating for several days and jumping in the same trip, a "normal" interstellar trip is acceleration for a few hours (normal 100 D limit), jump for a week, and acceleration for a few hours. Stellar 100 D limits were not considered in LBB2'77.
Or they just grabbed an arbitrary number that sounded good for small ships (10Pn) and another arbitrary number that seemed to work for smallcraft (10 kg/G/turn).
What they actually said is that ships use fuel to produce power that is fed to the "manoeuvre drive", but smallcraft use fuel directly in the "manoeuvre drive". To me (probably influenced by later editions) that does not sound like the same technology, more like magical power-to-thrust machines in ships, and magically fuel-efficient rockets in smallcraft. I suspect they started with rockets, but removed it as too much faff, replaced with "don't worry about it, it just works". It should probably not be taken all that literally, or assumed to be all that perfect: It's just a small part of a game.
In the game Imperium (developed concurrently with Traveller?) ships could accelerate to about 85% of lightspeed to travel between stellar systems, that would take a year or so of acceleration at 1 G. I would not take that to mean that all ships can routinely accelerate for years with standard amounts of fuel...