Some religious apologists will say that the fact that science "works" does not necessarily mean that it tells us anything true about the world. It seems to me that it is impossible to ever determine with absolute certainty what is true because what we see is a model of the world created by our brains. Moreover, the accuracy and precision of our senses even with the aid of instruments is limited. Absolute certainty is impossible. But this is not the standard we use to assess our world outside of mathematics. It is sufficient to speak of what we can assign a very high probability of being true. But let's consider what might happen were one to proceed only when absolutely certain of the veracity of a proposition, for example, "There is a wall here."
The evidence of my senses tells me there is a wall in front of me. I can believe that there is no wall because I know that empirical evidence does not necessarily tell us what is true. I can then attempt to walk through the space where my senses indicate there is a wall, but I will be unable to do so. My progress will be impeded by what I, having decided not to accept the evidence of my senses, can only describe as "something." But though I doubt my senses, I must now account for the problem of the "something" that is impeding my progress to the other room. I could assume a mysterious force prevents me passing, but I have no particular reason to believe this is true either. My senses tell me there is a wall, so I could go with the simplest assumption and decide that a wall accounts sufficiently for both my lack of progress and the phenomenon that is currently stimulating my visual cortex. I could venture that the two forms of sense data are correlated. I could then say that to the extent that different sense data are correlated they tell me something true about the external world.
But perhaps I decide that both sense data do not necessarily tell me what is true. How can I justify believing in evidence? So I go on stubbornly denying that the wall is truly there. I have no valid reason to believe that it is truly there. But now I have a problem. I am not succeeding in achieving my goal of reaching the next room by attempting to pass through the thing currently impeding my progress, and I decide to go around what my senses tell me is a wall. Going around the wall "worked." Or did it? Perhaps I never moved at all. I have only my senses to tell me that I did, and they do not necessarily tell me what is true. And the game goes on.
What will satisfy the theologians? What do they expect? How do we know science tells us anything true? Indeed, I could posit an omniscient god that knows what is true, but I don't see how this assumption helps. I would then need to explain how I could verify the assumption. How can the theist know with certainty that an omniscient deity exists and is not merely delusion without himself being omniscient?
Since none of us can know a mind-independent reality unfiltered by human brains, the logic of science is necessarily probabilistic. The theist can fairly say that science can not give us absolute truth, but then we have established that the process of filtering reality, whatever that may be, through human brains can by definition never give us direct knowledge of a mind-independent world. The theist faces the same difficulty. In this way, the theist and the naturalist are equal. But the crucial difference is that the method of science significantly increases the probability of achieving a given outcome. One could argue that this is because science maps onto reality. Of course as Hume pointed out, we have no reason to expect that because the sun has always risen in our experience that it will necessarily rise tomorrow. So science is relegated to the consistency of its results, to its predictive power. Science "works." Does the fact that it works increase the probability that the method of science tells us what is true? Does the astonishing accuracy of our mathematical models which undergird our theories increase the probability that they are true? We could after all be brains in vats, and what we perceive as reality could be a simulation. But if so, we could never know it. Our world, whether virtual or actual, is for our purposes, sufficiently real.