If you are drawing 40A, that's 40/3 Amps per battery. Each battery needs a fuse that can handle a little over 13A. The 1210 size fuses are better to use than the 1206 size. They're more robust and have lower resistance. I would suggest you use two of the 1210L450SLWR in parallel on each cell. They're quite small and should not be too big for you.
Mamu posted some pics somewhere where the fuses were mounted on a host board. It's really not something that requires a tutorial, not much too it. Just cut a small piece of copper clad board with a Dremel tool then a shallow cut down the middle to make two pads. Then solder the fuses and wires on to it.
For MOSFETs used as power switches, the only consideration really is the "on" resistance and gate-source threshold voltage. Typically the lower the "on" resistance the more current they can handle. "On" resistance and maximal current are usually specified at the top of the data sheet. For a set of parallel cells you also need a low gate-source threshold. You can spot a low gate-source threshold by an "on" resistance rating at 2.5V, again at the top of the data sheet.
MOSFETs come in a huge array of package types. I always use the SMD type so I don't look at the TO-220 and similar packages very often. I did see one linked below that looked good for a single cell or set of parallel cells. It's an N-channel so you need to switch the low side instead of the high side, but for an unregulated mod it will make no difference functionally.
http://www.fairchildsemi.com/ds/FD/FDD3706.pdfThe problem with the MOSFET you mentioned is it has a rather high "on" resistance so it has a lower current limit. It does have the 2.5V rating, but it will get pretty hot at 40A and waste considerable power. The one I linked to above has much lower "on" resistance. It will waste less power and run quite a bit cooler.
40A is a very large current demand. Ideally you don't want to insert any more than a few mOhms between the atomizer and battery. Consider power wasted is current squared times resistance so even an insertion of 5 mOhms will waste 8 Watts and drop two tenths of a Volt. Pretty shocking, but that's what happens with extreme currents.
The MOSFET I listed is still going to introduce a lot of loss. MOSFETs can be found with "on" resistance down to a couple three mOhms, but the 2.5V specified ones with "on" resistance that low are hard to find at best in a leaded package. If push comes to shove, you can parallel two MOSFETs to cut "on" resistance in half.
The fuses I mentioned should be fine in terms of power and voltage loss, you'll only get an mOhm or two in total with a pair on each cell.
Consider raising voltage rather than lowering resistance to obtain more power. It's a lot more efficient. The paradigm in vaping is to drop resistance for more power, but it's more efficient to increase voltage. You would actually do better with series cells powering series coils.
