Richard makes a good point here: first you'll have to make sure how noise is introduce before you will be able to reduce it to a usefull level. There is a major difference between conducted and radiated noise here!

Under all circumstances, a good PCB layout is the primary target. All these high-tech PCB routing programs create your PCB in no-time, but it my experience that most of the low priced (and even some of the higher priced such as Orcad!) don't take ground loops so serious. If you don't route the board manually, at least take your time in routing the power lines manually or carefully check for loops and correct as necessary.

Secondly, do not underestimate the design of you power supply. Making a regulated 5V supply with a standard 3-pin regulator appears not to be as easy as one might think; I've seen many supplies that run on the edge of oscillation due to incorrect capacitor sizing, so that they need next to nothing to tip over and produce the weirdest supply voltage you can imagine.

Furthermore take your time in designing I/O filters if you go directly into the field with processor ports. I wouldn't recommend to do that in the first place, but if you do consider adding some low-pass filters. Rember that each field cable will work as an antenna, so in the end you might end up building a radio receiver rather than a control system. You can experiment with some ferrites for HF suppression, but it is still better to separate your I/O signals galvanically, which may not be that difficult at all and may require nothing more than a hand full of opto-couplers.

Finally, put some small ceramic capacitors at strategic positions on your board to decouple your supply as close as possible to the supply pins of semiconductor devices. You'll be amazed what it does!