If you have 300 lead-acid cells in series then I have to assume that you have very big cells intended to drive a truck or similar. For lower current consumption, it would be better to have much fewer cells and instead step up the voltage than to stack huge numbers of cells.

So if you now have 300 very large cells, then the cost of each cell is also significant and you can afford a bit more fancy electronics.

Making a 300-cell multiplexer would not be a good idea for many reasons.

Why not see you 300 cells as 30 stacks of 10 cells each or maybe 25 12-cell stacks or whatever works well. For each stack, add one complete measuring module with a MUX, ADC and microcontroller. Then use opto-couplers to let a master processor get the readins digitally from the stacks.

Then you can handle 600V without any problems, and you can scale your design anyway you like - it would not matter if you have multiple stacks in series or in parallell. The master processor would still just collect readings from stack m, cell n, and decide of the individual cells behaves.

And as Andy have already noted, Linear have very nice chips than can handle a single stack of cells. The Linear chips have serial output and supports stacking, so all you need is to pick up the results for analysis. Your problem really gets much easier if you don't see 300 individual cells but instead think of m chains of n cells.