"The delta V to land on the moon is only slightly less than the Delta V to get to mars - where is the POSSIBLE sensibility in using the moon as a 'filling' station to go to Mars?"

It's about 2 km/sec more delta V to land on the Moon than to get to Mars. Though it is slightly less delta V to get to the Moon than to get to Mars.
In terms of a filling station for going to Mars, one should talk of delta V in regards to Earth/Moon L-points. If you speaking of Manned Mars, one has to consider travel time as a more important factor- sending a unmanned spacecraft to Mars, one could use a trajectory that take 9 months with a low delta V- you don't do this if you have crew.
With a crew going to Mars your possible travel time could be in the range of 1-6 months. 1-3 months travel time is not practical for chemical rockets- you looking a total delta V of over 20 km/sec from any earth orbit.

From LEO it takes about 3 km/sec to get to Earth escape or L-points; from a L-point you are about .05 km/sec delta V from an earth escape trajectory [or any other L-point] - or you require 3 km/sec less delta V to get to any destination beyond Earth as compared to starting from LEO. So if from LEO requiring say 12 km sec delta V to get to Mars, the same trajectory requires 9 km/sec from a L-point.
Earth/Moon L-1 is the nearest L-point to Earth. It is also the nearest L-point to the Moon.
Going to Mars from L-1, if you using LH&LOX [the best chemical rocket fuel in the space environmnet] and if your total Delta V is going to be around 9 km /sec, about 90% of the mass of this space craft is going to be rocket fuel.
[It takes about 9 km/sec delta V to get to LEO from Earth surface, rocket engines are less efficent in atmosphere, the rocket's structure needs to be strong enough to withstand the air pressure from high speed and the need for efficency of high gee acceleration. For instance crew of Shuttle experience 3 gees from acceleration plus the shock from solid boosters, so you looking at structure strong enough to handle all these loads {probably has a structural rating for *at least* 10 gees}. Whereas in the space environment one doesn't have this high of a requirement [for efficency sake] of high gee acceleration. 1-2 gees aceleration is probably enough so your overall structural strength could be quite safe at being half or 1/3 of the Shuttle or other space launchers. You have an optimal burn window at Earth's perigee of about 20 mins- about 4 times longer than a rocket launching from Earth.]

So, say the payload part that will be landing on Mars weighs 50 tons. The space vehicle at L-1 could have a mass of 500 tons with 400 tons of this mass never leaving the Earth system- it's a stage one of Mars manned space vehicle which gets refueled and reused. For this first stage the dry mass could be quite low- say less than 5%- less than 20 tons. And fuel burned for accelerating the second part of the stage could be over 350 tons with remaining fuel used to keep the first stage in the Earth system. So about 3/4 of way through the Perigee burn, the first separates, turns around and decelerates. The second stage part fires it's engines and consumes a portion of it's fuel, but leaving enough fuel so that it can decelerate to be captured in Mars orbit.
This Mars vehicle could have 3 stages, the booster that puts it on a Mars trajectory, a command module stage that adds to Mars trajectory and decelerate at Mars, this stage could also serve as long duration living quarters, have solar panels, water cycling, solar flare shelter, etc. And third stage is a lander- it lands on Mars and needs to be refueled to get back into orbit.

This is of course a rough sketch in acually designing you could probably reduce the mars lander gross weight, instead of 50 tons, one might get it down to 10-20 tons. The dry mass of the stage one might be 2%. And overall gross mass might be around 250 tons instead of 500 tons. Or you could want fast transition time to Mars if the premium is 4 month or less, you could looking at something grossing over 600 tons.

And there is not much sense going quickly to Mars and then returning to Earth at a much longer duration [other than you have medical attention available on Earth {and if you already have crew on Mars, they could assist newly arriving crew to Mars}. So what I'm discussing not only need a gas station at L-1, but also needs gas station in Mars orbit and Mars surface. In term of gross payload to Mars more than 90% can go to Mars using low delta-v and 9 month transition time. The express run is only used for crew or emergency resupply.

In terms of cost of all this fuel it's not that expensive in terms of the entire cost of a program. In quantities of hundreds of tons, a reasonable near term cost may be around $4000 per lb at L-1, so 1000 tons cost 8 billion. And in quantities of thousand tons the cost per lb could be around $2000 per lb. At Mars orbit the cost in quantities of hundreds of tons it could be around $5000 per lb. On Mars surface the main cost would be getting a nuclear reactor on the surface- solar panels are at least half as efficent on Mars as compared to Earth surface, and because of day and nite half again as efficent as to compared to L-1 or high orbits. Solar panels on Mars only makes sense when you want keep infrastructure costs low and have a low energy needs. With abundant electrical power from a nuclear reactor and with there being abundant quantities of water available, rocket fuel is pretty cheap on the Mars.

I would describe a very robust NASA Manned Mars of say 60 crew going to Mars over a period of 10 years, and the rocket fuel cost could over 50 billion, a more realistic Manned Mars might have half as many NASA crew over 10 year period and about 25 billion in rocket fuel cost, and this would be about the cost of ISS over a 10 year period. And if you add in all the other Manned Mars cost, you could it somewhere around what NASA is spending on the Shuttle program over a 10 year period.
In other words for a 20 year Mars Manned program you would have a cost of around 100 billion- what NASA is currently spending on Shuttle and ISS.