Rain1224, of course we could all look up everything we ever wanted to know if it has been done before and written about, and most folks try to do a bit of research before posting a question, but it is often more informative to discuss it amongst people doing similar things to get the best outcome. The question asked here is a good example, because you can get more than one conflicting answer out of different books. Discussing it among people doing similar things is a good way to come up with a reasonable solution. Besides you get to meet folks from all over the World here.
As for the question at hand:
The one modification I would make to your plan with the modules would be to measure the short circuit current and the open circuit voltage of each module under similar conditions. Any that are low on output that I believe would not reach sufficient voltage to contribute to the current I would exclude, say any under 8 volts under reasonable daylight, and then I would divide the remaining into two sets of modules which I would parallel such that each of those two groups put out approximately the same current.
Should there be three or more modules with low voltage I would hook those in series by three (or more) in a separate string if I could arrange it to reach charging voltage so that it would contribute current and then parallel that with the main group. Anyway, just my thoughts on it. Rich