Hello,
A google search, unsurprisingly, did not lead to any TLC prediction programs. It's probably impossible to model TLC accurately for practical application because there are many factors which contribute to rf values and separation which differ by things such as humidity, for example. Some of the factors are water on the plate and in the sample, the way the solid phase was laid on the plate, hydrogen bonding in the sample, amount spotted on a plate, residual solvents such as DMF, plate backing etc. My guess is a TLC modeling program, if used exclusively, will cause many columns to give poor separation.
If you are having trouble picking out solvents, it's best to try out a few. Eventually you get a feel for what solvents and in what ratios will move a compound to the rf you want.
Best