It is hard to say what to change when we cannot see how each program is handling the water volumes. Are the losses to trub different? Is the grain absorption different? Is the boil off rate different? is the batch size the same?
That said, the most important question is, "which prediction is closer to your actual results?"
The problem with trying to use two different models, even those built on the same basic equations, is that you will spend more time trying to get them to agree with each other than it will be worth in the long run. Pick one of the programs to use and concentrate on getting that program to closely match your actual results. THEN, if you want you can focus on the other program with a better understanding of the changes needed to make the first program mimic your process.