I can provide evidence that my memories are real, because I can use them to make accurate novel predictions that I can then test in reality. I can then have proportional confidence that my memories are true to the extent to which they can make these accurate predictions.
For example: If I remember leaving my jacket in the car instead of in my closet, and I go out to the car and find my jacket there, then my memory corresponded to reality quite accurately and was "true" in that sense.
Note the emphasis on "novel" predictions, meaning using your conceptual model of reality to extend it to things no one has tested yet. We can instead post hoc rationalize any explanation for something we already know, and make non-novel predictions, but this has very little value for explaining reality. For example, "If my claim that fairies exist is true, then we will observe the sun rising tomorrow morning," is not a novel prediction because we already know the sun rises every morning. The same goes for, "If invisible magical turtles exist who cause trees to grow, then we will see trees growing over time. Since trees do grow, this is evidence for invisible magical turtles who cause trees to grow."
In my experience, religious claims can only be supported by this kind of post hoc rationalization. "The universe exists, and this is evidence for a being that can create universes." I have never seen any theist use their theological model of reality to make an accurate novel testable prediction that was then tested and confirmed. Prayer studies are one example, where we would expect prayer to cause a statistically significant increase in good outcomes for healing, but studies reliably show its effect to be identical to the control group, meaning no effect.
If an idea cannot be extended to accurately model previously unobserved reality in this way, then it is very likely an imaginary idea. Memories do not fall into this imaginary category, by and large, whereas religious ideas apparently always do.