leeinmt, I am into FPV myself heading to the North American FPV meet in Montreal this weekend.
Anyway IMO two of the most important things to consider when building an ROV is:
1. Hull integrity. Design, build, test, test and retest
BEFORE you load it with electronics and batteries. If you build it correctly and know its limitations you should never have an issue. I know my design will operate all day at a 200 feet depth. It might even make it to 300 feet but my known testing limit is 200 feet so that's where I limit myself.
A simple way to test your design is to first select a target depth that you will typically operate your ROV at. Seal your hull, and throw some weight in it as if it was loaded with your onboard equipment. Now throw some external weight on the exterior and submerge it your desired working depth. Leave it there for an hour, two hours or overnight if you like. Return it to the surface, open and inspect. No leaks, no damage, then you know your design works and have somewhat figured out a safe margin to keep your dives at. Retest, its best you verify you have repeatable results. I personally tested my system about half a dozen times. Anchored to the bottom, with a buoy on the surface and left over night. If you have a failure then you need to redesign and the only thing lost is your time and a little material not a $1000 dollars worth of hardware. Its the KISS approach for pressure testing but it is a proven method.
2. Knowing what's going on inside. Like FPV get an OSD to display data on your screen. Monitor battery voltages, current consumption, internal/external pressures, temperature, compass headings etc... and most importantly a leak detection system. It goes without saying if you get so much as one drop of water inside you want to know about it as soon as possible, so a properly designed leak detection system is a must.
This is my homebrew system and it performs exceptionally well.
http://www.youtube.com/watch?v=moE6Oryyvww