Mars has two moons: Phobos and Deimos. Both are irregular and are believed to have been captured from the nearby asteroid belt.
Phobos always shows the same face to Mars because of tidal forces exerted by the planet on its satellite. These same forces causes Phobos to drift increasingly closer to Mars, a situation that will cause their collision in about 50 to 100 million years.
How I can calculate, given appropriate data, the estimated time at which Phobos will collide with Mars?
First, you state a few things that aren't quite right in your question. While the view that's generally talked about is that Phobos and Deimos are likely captured asteroids, dynamically it's a pretty difficult problem (you generally need a third (in this case fourth?) body to take away the extra energy, and it's hard to get a circular orbit around the equator). See for a bit more on that.
In terms of Phobos' demise, there are two things that make this problem very difficult to estimate. First, Phobos' orbit evolves as it orbits around Mars, so you can't just take a linear approach and say, "It's moving towards Mars at 18.3 cm/year so it's going to hit in about 50 million years." It's more complicated and non-linear.
But besides that, there's the Roche Limit to consider, whereby the moon will break up due to tidal forces before it would actually hit. The problem there is that Phobos is already within the Roche Limit, meaning that it's only being held together now by the physical strength of the rock it's made of. And since we don't really know what it's made of inside (though we can make educated guesses and I'm sure there are models out there for its strength), these unknowns make it somewhat difficult to estimate.
Another symptom of Phobos being so close to mars is that objects on its surface are not all in zero G. On average, object at the same altitude will weigh about 0.285 kg*mm/s less per kg on the day side than on the twilight one, with a gravity sometimes dipping under 2 mm/s^2, this could have noticeable effect on escape velocity.
Additionally, Phobos escape is not really necessary to escape Phobos, a jump above Stickney crater that would launch you all of 3 centimeters up on Earth would be enough to get to the Lagrange L1 point between Phobos and Mars. So you would than be in zero G, and never fall back down (staying in mars orbit) (although this jump would take around half an hour on the way up).
So, an astronaut trying to move around on said crater would be nightmarish without some sort of rocket pack, Let's say he bends over to pick up a rock, he gets back up over the course of 2 seconds, moving his center of mass 1 meter, he will now be moving a minimum of 1/2 meter/s away from Phobos, easily enough to reach L1 and go into a Mars orbit. Let's say that (s)he tries to walk quickly, his/her center of mass moves at several tenths of a meter per second upward, in addition to his m/s forward, therefore, he moves upward and drifts again into Mars orbit. Throw a 5 kg rock downward? go into orbit, throw it upward? still go into orbit, as does the rock, Here is the scary part, people on Phobos might lift not just that rock, but there lander if they try very hard at all. Oh, and heaven help you if you thought firing a gun downward was a good idea! less than a half-second of assault-rifle rounds into the crater will send you to Mars orbit. Or what if you puncture your suit? You might think you could just hurry up and be inside your pressurized spaceship within 2 minutes and you'd be fine, but nope, if it is facing upwards, air will rush out at 340 m/s, if it is escaping through a 1 cm^2 hole (widened by the pressure), you would lose 34 liters per second of air at 1/3rd sea-level pressure, even if your tank holds hundreds of times that, you still emit about 12 grams/s of air, accelerating you at 0.0408 m/s^2, enough to get to get to L1 with around 7 seconds of air released.
Oh, and don't even think about using golf to demo low gravity. 2 kg of metal flying back up at an appropriate speed could Easily send you into mars orbit.
As could a poorly balanced washing machine (send anything on top of it).
Diet Coke and Mentos rockets could reach escape velocity. Forget diet coke and Mentos, Diet coke by itself might be able to Reach L1.
It would take only a few nukes to deorbit Phobos into mars, creating an explosion that obliterates whole martian regions.
Dynamical models over a likely timescale (say $10^6$ to $10^9$ years) would have significant error bars as mentioned above, and therefore one off predictions about individual moons have little validity.
Because Phobos is already within the Roche Limit, within which it should disintegrate due to tidal forces, "appropriate data" would have to include quite a bit of detail about Phobos's structure and composition (information which we currently lack), which would let you determine not so much the "time at which" it will collide, but the period over which bits of Phobos would.
